-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  C:\Users\knm4642\Dropbox\Nha Meo Heo Beo\_JMP\WORKING FOLDER\Table A6_Robustness checks.log
  log type:  text
 opened on:  18 Feb 2026, 23:49:43

. 
. //PREPARE DATA
. use "Baseline CEO firm year sample.dta", clear

. * set globals
. global firm = "firmage firmage2"

. global ceo = "gender age age2 yrinco i.education"

. global sample = "year < 2012 & !nonUS"

. global cluster = "mainethcode"

. * set sample
. qui reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)

. gen sample = e(sample)

. 
. 
. //PREPARE TABLE
. * PANEL A: Alternative control variables and Poisson model
. eststo clear

. eststo col1: /// no controls
>         reghdfe ash_f1allpat trust_sd if sample, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 6 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(   1,     37) =      10.65
Statistics robust to heteroskedasticity           Prob > F        =     0.0024
                                                  R-squared       =     0.9012
                                                  Adj R-squared   =     0.8873
                                                  Within R-sq.    =     0.0005
Number of clusters (mainethcode) =         38     Root MSE        =     0.5430

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0588299   .0180304     3.26   0.002     .0222968     .095363
       _cons |   .7030451   .0908329     7.74   0.000     .5190001    .8870901
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "No ctrls"

added macro:
               e(spec) : "No ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col1, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col2: /// baseline
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      20.26
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0022
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0618858   .0168605     3.67   0.001     .0277232    .0960484
     firmage |  -.0109686   .0019023    -5.77   0.000    -.0148231   -.0071142
    firmage2 |   .0001159   .0000492     2.35   0.024     .0000162    .0002157
      gender |  -.0118746    .033103    -0.36   0.722    -.0789476    .0551983
         age |  -.0012187   .0103035    -0.12   0.906    -.0220956    .0196582
        age2 |  -6.07e-06   .0000846    -0.07   0.943    -.0001776    .0001654
      yrinco |   .0040659   .0005505     7.39   0.000     .0029505    .0051814
             |
   education |
          2  |   .0426374   .0480788     0.89   0.381    -.0547795    .1400544
          3  |   .0592057      .0333     1.78   0.084    -.0082666     .126678
          4  |   .0837448   .0411572     2.03   0.049     .0003525    .1671372
             |
       _cons |   .8168338   .2952368     2.77   0.009     .2186271     1.41504
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Baseline"

added macro:
               e(spec) : "Baseline"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col2, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col3: /// logat control
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} logat if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  11,     37) =      75.73
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9033
                                                  Adj R-squared   =     0.8897
                                                  Within R-sq.    =     0.0215
Number of clusters (mainethcode) =         38     Root MSE        =     0.5374

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0626851   .0187413     3.34   0.002     .0247116    .1006587
     firmage |  -.0005056   .0018575    -0.27   0.787    -.0042693    .0032581
    firmage2 |   .0001983   .0000469     4.22   0.000     .0001031    .0002934
      gender |  -.0033152   .0313104    -0.11   0.916     -.066756    .0601257
         age |  -.0077657   .0105938    -0.73   0.468    -.0292308    .0136994
        age2 |   .0000542   .0000873     0.62   0.539    -.0001228    .0002312
      yrinco |   .0030431   .0005959     5.11   0.000     .0018357    .0042505
             |
   education |
          2  |   .0309306    .048858     0.63   0.531     -.068065    .1299262
          3  |   .0470342   .0338222     1.39   0.173    -.0214961    .1155645
          4  |   .0649291   .0425408     1.53   0.135    -.0212669     .151125
             |
       logat |   .1600797    .011681    13.70   0.000     .1364117    .1837477
       _cons |  -.2116385   .3079308    -0.69   0.496    -.8355656    .4122886
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Add ctrls"

added macro:
               e(spec) : "Add ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col3, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col4: /// logsale control
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} logsale if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  11,     37) =      70.87
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9020
                                                  Adj R-squared   =     0.8883
                                                  Within R-sq.    =     0.0091
Number of clusters (mainethcode) =         38     Root MSE        =     0.5408

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0616875   .0179651     3.43   0.001     .0252868    .0980882
     firmage |  -.0079321   .0016328    -4.86   0.000    -.0112405   -.0046237
    firmage2 |   .0001908   .0000457     4.17   0.000     .0000981    .0002835
      gender |  -.0100629    .031971    -0.31   0.755    -.0748423    .0547164
         age |  -.0052764   .0103382    -0.51   0.613    -.0262236    .0156708
        age2 |   .0000294   .0000847     0.35   0.731    -.0001423    .0002011
      yrinco |   .0037473    .000612     6.12   0.000     .0025072    .0049874
             |
   education |
          2  |   .0374127   .0473161     0.79   0.434    -.0584588    .1332841
          3  |   .0542648   .0325667     1.67   0.104    -.0117216    .1202512
          4  |    .078931   .0405428     1.95   0.059    -.0032166    .1610786
             |
     logsale |   .0794696   .0096062     8.27   0.000     .0600056    .0989336
       _cons |   .3780847   .3430199     1.10   0.277    -.3169396    1.073109
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Add ctrls"

added macro:
               e(spec) : "Add ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col4, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col5: /// logemp control
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} logemp if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     28,506
Absorbing 2 HDFE groups                           F(  11,     37) =     153.85
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9044
                                                  Adj R-squared   =     0.8908
                                                  Within R-sq.    =     0.0201
Number of clusters (mainethcode) =         38     Root MSE        =     0.5380

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0652415   .0184223     3.54   0.001     .0279144    .1025685
     firmage |  -.0048896   .0022195    -2.20   0.034    -.0093866   -.0003925
    firmage2 |    .000243   .0000457     5.31   0.000     .0001503    .0003356
      gender |  -.0089499   .0336822    -0.27   0.792    -.0771965    .0592967
         age |  -.0102852   .0102765    -1.00   0.323    -.0311073    .0105369
        age2 |   .0000714   .0000842     0.85   0.402    -.0000992    .0002421
      yrinco |   .0031144   .0007266     4.29   0.000     .0016421    .0045866
             |
   education |
          2  |   .0582232   .0554218     1.05   0.300    -.0540721    .1705184
          3  |   .0730926   .0402812     1.81   0.078    -.0085248    .1547101
          4  |   .0911282   .0468817     1.94   0.060    -.0038632    .1861196
             |
      logemp |    .180855   .0096839    18.68   0.000     .1612336    .2004764
       _cons |   .8228276   .3225284     2.55   0.015     .1693229    1.476332
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3548           0        3548     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Add ctrls"

added macro:
               e(spec) : "Add ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      28506       3548

.                 eststo col5, add(nofirms r(ndistinct))
(e(nofirms) = 3548 added)

. eststo col6: /// asinh(rdstock) control
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} ash_rdstock if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  11,     37) =      33.66
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9015
                                                  Adj R-squared   =     0.8877
                                                  Within R-sq.    =     0.0041
Number of clusters (mainethcode) =         38     Root MSE        =     0.5421

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0628148   .0169568     3.70   0.001      .028457    .0971725
     firmage |  -.0135172   .0016165    -8.36   0.000    -.0167926   -.0102418
    firmage2 |   .0001598   .0000468     3.42   0.002     .0000651    .0002545
      gender |  -.0091182   .0328627    -0.28   0.783    -.0757043    .0574679
         age |  -.0033661   .0106828    -0.32   0.754    -.0250115    .0182792
        age2 |   .0000124   .0000882     0.14   0.889    -.0001663    .0001912
      yrinco |   .0040215   .0005547     7.25   0.000     .0028976    .0051455
             |
   education |
          2  |   .0388324    .047967     0.81   0.423     -.058358    .1360228
          3  |    .054171   .0336763     1.61   0.116    -.0140636    .1224056
          4  |   .0803669   .0418939     1.92   0.063    -.0045181    .1652519
             |
 ash_rdstock |   .0617015   .0138292     4.46   0.000     .0336809     .089722
       _cons |   .7344658   .3018132     2.43   0.020     .1229342    1.345997
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Add ctrls"

added macro:
               e(spec) : "Add ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col6, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col7: /// asinh(xrd) control
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} ash_xrd if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  11,     37) =      28.91
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9026
                                                  Adj R-squared   =     0.8890
                                                  Within R-sq.    =     0.0152
Number of clusters (mainethcode) =         38     Root MSE        =     0.5391

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |    .058135   .0166883     3.48   0.001     .0243213    .0919487
     firmage |  -.0099808   .0018964    -5.26   0.000    -.0138234   -.0061383
    firmage2 |   .0001367   .0000477     2.86   0.007       .00004    .0002334
      gender |  -.0153965   .0344485    -0.45   0.658    -.0851959    .0544029
         age |  -.0033118   .0101838    -0.33   0.747    -.0239461    .0173225
        age2 |   .0000141   .0000843     0.17   0.868    -.0001566    .0001848
      yrinco |   .0034183   .0005786     5.91   0.000     .0022459    .0045907
             |
   education |
          2  |   .0389201   .0460243     0.85   0.403    -.0543339    .1321741
          3  |   .0515698   .0319673     1.61   0.115     -.013202    .1163417
          4  |   .0761822    .038795     1.96   0.057    -.0024238    .1547883
             |
     ash_xrd |   .1383377   .0147939     9.35   0.000     .1083625     .168313
       _cons |   .6156426   .3016483     2.04   0.048     .0044451     1.22684
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local spec       "Add ctrls"

added macro:
               e(spec) : "Add ctrls"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col7, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. xtset boardid

Panel variable: boardid (unbalanced)

. eststo col8: /// poisson
>         xtpoisson f1allpat trust_sd ${firm} ${ceo} i.year if ${sample}, fe vce(robust) 
note: 189 groups (189 obs) dropped because of only one obs per group
note: 1494 groups (11659 obs) dropped because of all zero outcomes

Iteration 0:  Log pseudolikelihood =  -69131.88  
Iteration 1:  Log pseudolikelihood = -64169.134  
Iteration 2:  Log pseudolikelihood = -64150.291  
Iteration 3:  Log pseudolikelihood = -64150.289  

Conditional fixed-effects Poisson regression         Number of obs    = 17,536
Group variable: boardid                              Number of groups =  1,915

                                                     Obs per group:
                                                                  min =      2
                                                                  avg =    9.2
                                                                  max =     20

                                                     Wald chi2(21)    = 177.39
Log pseudolikelihood = -64150.289                    Prob > chi2      = 0.0000

                                (Std. err. adjusted for clustering on boardid)
------------------------------------------------------------------------------
             |               Robust
    f1allpat | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .2516414   .0771332     3.26   0.001     .1004631    .4028196
     firmage |   .0216057   .0142252     1.52   0.129    -.0062752    .0494866
    firmage2 |  -.0004295   .0002333    -1.84   0.066    -.0008869    .0000278
      gender |    .005622   .1024483     0.05   0.956     -.195173     .206417
         age |   .0159146   .0580795     0.27   0.784    -.0979192    .1297484
        age2 |  -.0002148   .0005206    -0.41   0.680    -.0012351    .0008056
      yrinco |   .0120841   .0029624     4.08   0.000     .0062779    .0178904
             |
   education |
          2  |   .6186062   .1270686     4.87   0.000     .3695563     .867656
          3  |   .6894762   .1360182     5.07   0.000     .4228854     .956067
          4  |   .6667607   .1585785     4.20   0.000     .3559525    .9775689
             |
        year |
       2001  |   .0382628   .0390634     0.98   0.327    -.0383001    .1148258
       2002  |   .0405248   .0545962     0.74   0.458    -.0664817    .1475313
       2003  |   .0673916   .0722425     0.93   0.351    -.0742012    .2089844
       2004  |   .1507708    .093345     1.62   0.106     -.032182    .3337236
       2005  |   .1522197   .0951026     1.60   0.109    -.0341779    .3386174
       2006  |   .2132995   .1060013     2.01   0.044     .0055409    .4210582
       2007  |   .2064056   .1242239     1.66   0.097    -.0370688      .44988
       2008  |   .0236171   .1335495     0.18   0.860    -.2381352    .2853694
       2009  |   .0272891   .1455142     0.19   0.851    -.2579134    .3124917
       2010  |   .1575996   .1522158     1.04   0.300    -.1407379    .4559371
       2011  |    .290281   .1731865     1.68   0.094    -.0491583    .6297202
------------------------------------------------------------------------------

.                 sum f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
    f1allpat |     17,536    30.14137    191.1977          0       6875

.                 estadd scalar depvarmean = r(mean)

added scalar:
         e(depvarmean) =  30.141366

.                 estadd local spec       "Poisson"

added macro:
               e(spec) : "Poisson"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      17536       1915

.                 eststo col8, add(nofirms r(ndistinct))
(e(nofirms) = 1915 added)

. 
. global coeflist = "trust_sd logat logsale logemp ash_rdstock ash_xrd"

. esttab /*using "Table A6_Robustness checks_Panel A.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(18) modelwidth(12) ///
>         mgroups("Future patent applications", pattern(1 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(${coeflist}) order(${coeflist}) ///
>         coeflab(trust_sd "CEO's trust" ///
>                         logat "log(total assets)" ///
>                         logsale "log(sales)" ///
>                         logemp "log(employment)" ///
>                         ash_rdstock "arsinh(R\&D stock)" ///
>                         ash_xrd "arsinh(R\&D exp.)") ///
>         stats(spec FE controls N nofirms, fmt(%9.0fc %9.0fc) ///
>                 lab("Specification" "Firm \& Year FEs" "Baseline controls" "Observations" "Firms"))

--------------------------------------------------------------------------------------------------------------------------------------------------
                   \multicolumn{8}{c}{Future patent applications}                                                                                 
                            (1)             (2)             (3)             (4)             (5)             (6)             (7)             (8)   
--------------------------------------------------------------------------------------------------------------------------------------------------
main                                                                                                                                              
CEO's trust               0.059***        0.062***        0.063***        0.062***        0.065***        0.063***        0.058***        0.252***
                        (0.018)         (0.017)         (0.019)         (0.018)         (0.018)         (0.017)         (0.017)         (0.077)   
log(total assets)                                         0.160***                                                                                
                                                        (0.012)                                                                                   
log(sales)                                                                0.079***                                                                
                                                                        (0.010)                                                                   
log(employment)                                                                           0.181***                                                
                                                                                        (0.010)                                                   
arsinh(R\&D stock)                                                                                        0.062***                                
                                                                                                        (0.014)                                   
arsinh(R\&D exp.)                                                                                                         0.138***                
                                                                                                                        (0.015)                   
--------------------------------------------------------------------------------------------------------------------------------------------------
Specification          No ctrls        Baseline       Add ctrls       Add ctrls       Add ctrls       Add ctrls       Add ctrls         Poisson   
Firm \& Year FEs              X               X               X               X               X               X               X               X   
Baseline controls                             X               X               X               X               X               X               X   
Observations             29,384          29,384          29,384          29,384          28,506          29,384          29,384          17,536   
Firms                     3,598           3,598           3,598           3,598           3,548           3,598           3,598           1,915   
--------------------------------------------------------------------------------------------------------------------------------------------------

. 
. 
. * PANEL B: Alternative clustering schemes, weighting schemes, and patent transformations
. eststo clear

. cap drop TRUST

. gen TRUST = trust_sd 
(10,042 missing values generated)

. set seed 7

. eststo col1: /// Wild bootstrap
>          wildbootstrap areg ash_f1allpat TRUST ${firm} ${ceo} i.year if ${sample}, ///
>                 absorb(boardid) cluster(${cluster}) coefficients(TRUST)

Performing 1,000 replications for p-value for constraint
  TRUST = 0 ...
Computing confidence interval for TRUST
  Lower bound: .........10.........20....... done (27)
  Upper bound: .........10.........20. done (21)

Wild cluster bootstrap                            Number of obs      = 29,384
Linear regression, absorbing indicators           Number of clusters =     38
                                                  Cluster size:
Cluster variable: mainethcode                                    min =      3
Error weight: Rademacher                                         avg =  773.3
                                                                 max =   6027
-----------------------------------------------------------------------------
            ash_f1allpat |   Estimate      t  p-value    [95% conf. interval]
-------------------------+---------------------------------------------------
constraint               |
               TRUST = 0 |   .0618858    3.67   0.014    .0236883    .0950213
-----------------------------------------------------------------------------

.                 sum ash_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
ash_f1allpat |     29,384    1.001382    1.617756          0   9.528794

.                 estadd local cluster    "Wild"

added macro:
            e(cluster) : "Wild"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col1, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col2: /// cluster by CEO
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(ceoid) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,   5752) =       4.47
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0022
Number of clusters (ceoid)   =      5,753         Root MSE        =     0.5426

                              (Std. err. adjusted for 5,753 clusters in ceoid)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0618858   .0199363     3.10   0.002      .022803    .1009685
     firmage |  -.0109686   .0024898    -4.41   0.000    -.0158496   -.0060877
    firmage2 |   .0001159   .0000581     2.00   0.046     2.12e-06    .0002297
      gender |  -.0118746   .0372465    -0.32   0.750    -.0848919    .0611426
         age |  -.0012187   .0077377    -0.16   0.875    -.0163875    .0139502
        age2 |  -6.07e-06    .000069    -0.09   0.930    -.0001413    .0001292
      yrinco |   .0040659   .0009593     4.24   0.000     .0021854    .0059465
             |
   education |
          2  |   .0426374   .0298622     1.43   0.153    -.0159038    .1011786
          3  |   .0592057   .0309152     1.92   0.056    -.0013997    .1198112
          4  |   .0837448   .0343924     2.43   0.015     .0163227    .1511669
             |
       _cons |   .8168338   .2363624     3.46   0.001     .3534744    1.280193
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 sum ash_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
ash_f1allpat |     29,384    1.001382    1.617756          0   9.528794

.                 estadd local cluster    "CEO"

added macro:
            e(cluster) : "CEO"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col2, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col3: /// two-way clustering     
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(mainethcode boardid) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)
Warning: VCV matrix was non-positive semi-definite; adjustment from Cameron, Gelbach & Miller applied.

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =       8.42
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
Number of clusters (mainethcode) =         38     Within R-sq.    =     0.0022
Number of clusters (boardid) =      3,598         Root MSE        =     0.5426

                   (Std. err. adjusted for 38 clusters in mainethcode boardid)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0618858    .021048     2.94   0.006     .0192385    .1045331
     firmage |  -.0109686   .0021853    -5.02   0.000    -.0153964   -.0065409
    firmage2 |   .0001159   .0000607     1.91   0.064    -7.12e-06     .000239
      gender |  -.0118746   .0402655    -0.29   0.770    -.0934604    .0697111
         age |  -.0012187   .0102114    -0.12   0.906     -.021909    .0194716
        age2 |  -6.07e-06   .0000843    -0.07   0.943     -.000177    .0001648
      yrinco |   .0040659   .0008277     4.91   0.000     .0023889     .005743
             |
   education |
          2  |   .0426374   .0471414     0.90   0.372    -.0528801     .138155
          3  |   .0592057   .0343339     1.72   0.093    -.0103614    .1287729
          4  |   .0837448   .0425545     1.97   0.057    -.0024787    .1699683
             |
       _cons |   .8168338   .3001749     2.72   0.010     .2086217    1.425046
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598        3598           0    *|
        year |        12           1          11     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

.                 sum ash_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
ash_f1allpat |     29,384    1.001382    1.617756          0   9.528794

.                 estadd local cluster    "Two-way"

added macro:
            e(cluster) : "Two-way"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col3, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col4: /// aw = ethHHI
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample} [aw = ethHHI], ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      16.55
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9006
                                                  Adj R-squared   =     0.8866
                                                  Within R-sq.    =     0.0025
Number of clusters (mainethcode) =         38     Root MSE        =     0.5447

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0648217   .0185337     3.50   0.001     .0272688    .1023747
     firmage |  -.0103845   .0028092    -3.70   0.001    -.0160765   -.0046925
    firmage2 |   .0000995   .0000652     1.53   0.135    -.0000326    .0002317
      gender |   .0121159    .027203     0.45   0.659    -.0430026    .0672344
         age |  -.0015433   .0120475    -0.13   0.899    -.0259538    .0228672
        age2 |  -2.10e-06   .0001008    -0.02   0.984    -.0002064    .0002022
      yrinco |    .004794     .00071     6.75   0.000     .0033553    .0062326
             |
   education |
          2  |   .0636654   .0483567     1.32   0.196    -.0343145    .1616454
          3  |   .0842435    .035675     2.36   0.024     .0119591    .1565278
          4  |   .1026501   .0456675     2.25   0.031     .0101189    .1951813
             |
       _cons |    .761996   .3163699     2.41   0.021     .1209697    1.403022
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local weight     "Mapping precsion"

added macro:
             e(weight) : "Mapping precsion"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col4, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trustmaineth_sd
(36,055 real changes made, 10,412 to missing)

. eststo col5: /// trustmaineth_sd as X variable
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample} [aw = mainethshare], ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      17.44
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9003
                                                  Adj R-squared   =     0.8863
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         38     Root MSE        =     0.5447

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0545904   .0173408     3.15   0.003     .0194547    .0897261
     firmage |  -.0103922   .0024889    -4.18   0.000    -.0154352   -.0053492
    firmage2 |   .0001046   .0000604     1.73   0.092    -.0000177    .0002269
      gender |    .009027   .0289427     0.31   0.757    -.0496165    .0676705
         age |  -.0010037    .011412    -0.09   0.930    -.0241265    .0221192
        age2 |  -4.80e-06   .0000949    -0.05   0.960     -.000197    .0001874
      yrinco |   .0045337   .0006385     7.10   0.000       .00324    .0058273
             |
   education |
          2  |   .0615454   .0482962     1.27   0.210     -.036312    .1594028
          3  |   .0795639   .0347138     2.29   0.028     .0092271    .1499007
          4  |    .098183   .0444139     2.21   0.033     .0081918    .1881742
             |
       _cons |   .7937955   .3172957     2.50   0.017     .1508934    1.436698
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local weight     "Main ethnicity"

added macro:
             e(weight) : "Main ethnicity"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col5, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trust_sd
(36,055 real changes made)

. gen log1_f1allpat = log(1 + f1allpat)
(5,427 missing values generated)

. eststo col6: /// log1_f1allpat as Y variable
>         reghdfe log1_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      17.46
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9095
                                                  Adj R-squared   =     0.8968
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         38     Root MSE        =     0.4434

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
log1_f1all~t | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0526565   .0137077     3.84   0.000      .024882     .080431
     firmage |  -.0085154    .001574    -5.41   0.000    -.0117047   -.0053261
    firmage2 |   .0000778   .0000407     1.91   0.063    -4.59e-06    .0001602
      gender |  -.0139689   .0265527    -0.53   0.602    -.0677699     .039832
         age |  -.0013156   .0084661    -0.16   0.877    -.0184695    .0158383
        age2 |  -3.74e-06   .0000692    -0.05   0.957    -.0001439    .0001364
      yrinco |   .0036103   .0004513     8.00   0.000      .002696    .0045246
             |
   education |
          2  |   .0359027   .0385394     0.93   0.358    -.0421855    .1139909
          3  |   .0489574   .0266161     1.84   0.074     -.004972    .1028868
          4  |   .0692006   .0338716     2.04   0.048     .0005703    .1378308
             |
       _cons |   .6765262   .2429545     2.78   0.008     .1842537    1.168799
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "ln(1+.)"

added macro:
          e(transform) : "ln(1+.)"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col6, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col7: /// f1allpatw as Y variable
>         reghdfe f1allpatw TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      50.07
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9358
                                                  Adj R-squared   =     0.9268
                                                  Within R-sq.    =     0.0056
Number of clusters (mainethcode) =         38     Root MSE        =    17.3501

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
   f1allpatw | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   1.454633   .5589363     2.60   0.013     .3221206    2.587146
     firmage |   .0993034   .1040664     0.95   0.346    -.1115552    .3101621
    firmage2 |  -.0071851   .0032497    -2.21   0.033    -.0137697   -.0006005
      gender |   -2.45898   .8627323    -2.85   0.007    -4.207042   -.7109188
         age |   .1823269    .172871     1.05   0.298     -.167943    .5325968
        age2 |  -.0022466   .0016296    -1.38   0.176    -.0055486    .0010553
      yrinco |   .2109958   .0488929     4.32   0.000     .1119294    .3100622
             |
   education |
          2  |   .9320429   .5664909     1.65   0.108    -.2157766    2.079862
          3  |   1.996016    .969415     2.06   0.047     .0317943    3.960237
          4  |   1.373852   1.363028     1.01   0.320    -1.387905    4.135609
             |
       _cons |   3.751223   5.992572     0.63   0.535    -8.390881    15.89333
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 sum f1allpatw if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
   f1allpatw |     29,384    13.27883    64.14359          0        634

.                 estadd scalar depvarmean = r(mean)

added scalar:
         e(depvarmean) =  13.278825

.                 estadd local transform  "win"

added macro:
          e(transform) : "win"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col7, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col8: /// f1allpat as Y var
>         reghdfe f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      50.47
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9227
                                                  Adj R-squared   =     0.9118
                                                  Within R-sq.    =     0.0031
Number of clusters (mainethcode) =         38     Root MSE        =    44.0857

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
    f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   4.453607   1.240222     3.59   0.001     1.940677    6.966536
     firmage |   .0883571   .1362407     0.65   0.521    -.1876927    .3644069
    firmage2 |  -.0091204   .0036691    -2.49   0.018    -.0165547   -.0016861
      gender |  -1.538856   1.456126    -1.06   0.297    -4.489248    1.411537
         age |   .4052252   1.128543     0.36   0.722    -1.881421    2.691871
        age2 |  -.0058939   .0104664    -0.56   0.577    -.0271009     .015313
      yrinco |   .4079926   .1327433     3.07   0.004     .1390291    .6769561
             |
   education |
          2  |    11.4691   5.989296     1.91   0.063     -.666368    23.60457
          3  |   12.64381   4.947948     2.56   0.015     2.618313     22.6693
          4  |   11.73707   5.564751     2.11   0.042      .461816    23.01233
             |
       _cons |  -20.04418    31.2509    -0.64   0.525    -83.36451    43.27616
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 sum f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
    f1allpat |     29,384    18.01814    148.4439          0       6875

.                 estadd scalar depvarmean = r(mean)

added scalar:
         e(depvarmean) =  18.018139

.                 estadd local transform  "none"

added macro:
          e(transform) : "none"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col8, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. 
. esttab /*using "Table A6_Robustness checks_Panel B.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(18) modelwidth(12) ///
>         mgroups("Future patent applications", pattern(1 0 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(TRUST) order(TRUST) coeflab(TRUST "CEO's trust") ///
>         stats(depvarmean cluster weight transform FE controls N nofirms, fmt(%9.3fc %9.0fc) ///
>                 lab("Dep. var. mean" "Clustering scheme" "Weighting scheme" "Transformation" ///
>                         "Firm \& Year FEs" "Baseline controls" "Observations" "Firms"))

--------------------------------------------------------------------------------------------------------------------------------------------------
                   \multicolumn{8}{c}{Future patent applications}                                                                                 
                            (1)             (2)             (3)             (4)             (5)             (6)             (7)             (8)   
--------------------------------------------------------------------------------------------------------------------------------------------------
CEO's trust               0.062***        0.062***        0.062***        0.065***        0.055***        0.053***        1.455**         4.454***
                        (0.017)         (0.020)         (0.021)         (0.019)         (0.017)         (0.014)         (0.559)         (1.240)   
--------------------------------------------------------------------------------------------------------------------------------------------------
Dep. var. mean                                                                                                           13.279          18.018   
Clustering scheme          Wild             CEO         Two-way                                                                                   
Weighting scheme                                                   Mapping precsion    Main ethnicity                                                   
Transformation                                                                                          ln(1+.)             win            none   
Firm \& Year FEs              X               X               X               X               X               X               X               X   
Baseline controls             X               X               X               X               X               X               X               X   
Observations             29,384          29,384          29,384          29,384          29,384          29,384          29,384          29,384   
Firms                     3,598           3,598           3,598           3,598           3,598           3,598           3,598           3,598   
--------------------------------------------------------------------------------------------------------------------------------------------------

. 
. 
. * PANEL C: Alternative patent and trust measures
. eststo clear

. cap drop TRUST

. gen TRUST = trust_sd
(10,042 missing values generated)

. eststo col1: /// granted USPTO patents
>         reghdfe ash_f1uspat_granted TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      30.83
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9005
                                                  Adj R-squared   =     0.8865
                                                  Within R-sq.    =     0.0021
Number of clusters (mainethcode) =         38     Root MSE        =     0.4889

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1uspa~d | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0550037   .0135629     4.06   0.000     .0275227    .0824847
     firmage |  -.0138735   .0035764    -3.88   0.000    -.0211201   -.0066269
    firmage2 |   .0001093   .0000423     2.59   0.014     .0000237    .0001949
      gender |  -.0139129   .0184962    -0.75   0.457    -.0513897    .0235639
         age |  -.0024018   .0084195    -0.29   0.777    -.0194612    .0146577
        age2 |   .0000133   .0000706     0.19   0.851    -.0001298    .0001564
      yrinco |   .0031692   .0004228     7.50   0.000     .0023126    .0040258
             |
   education |
          2  |   .0027877   .0576461     0.05   0.962    -.1140143    .1195898
          3  |    .014275   .0475825     0.30   0.766    -.0821363    .1106862
          4  |   .0480403   .0611697     0.79   0.437    -.0759012    .1719818
             |
       _cons |   .7466277   .2183852     3.42   0.002     .3041373    1.189118
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local patent     "Granted USPTO"

added macro:
             e(patent) : "Granted USPTO"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col1, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trustlasso_sd
(46,825 real changes made, 14,239 to missing)

. eststo col2: /// LASSO-based trust
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      20.19
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9014
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0717814   .0216842     3.31   0.002     .0278451    .1157177
     firmage |  -.0111655   .0019291    -5.79   0.000    -.0150743   -.0072567
    firmage2 |   .0001148   .0000495     2.32   0.026     .0000146     .000215
      gender |   .0014111   .0337733     0.04   0.967    -.0670202    .0698424
         age |  -.0049813   .0111081    -0.45   0.656    -.0274885     .017526
        age2 |   .0000119   .0000885     0.13   0.894    -.0001675    .0001913
      yrinco |   .0038519   .0005434     7.09   0.000     .0027509    .0049529
             |
   education |
          2  |  -.0547697   .0725097    -0.76   0.455    -.2016883     .092149
          3  |   -.109158   .0761861    -1.43   0.160    -.2635258    .0452097
          4  |  -.1099224   .0862849    -1.27   0.211    -.2847523    .0649075
             |
       _cons |     1.0543   .3115606     3.38   0.002     .4230182    1.685582
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "LASSO"

added macro:
              e(trust) : "LASSO"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col2, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trustfullgss_sd
(46,825 real changes made)

. eststo col3: /// full-GSS-based trust
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      21.73
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8874
                                                  Within R-sq.    =     0.0018
Number of clusters (mainethcode) =         38     Root MSE        =     0.5427

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |    .038288   .0164728     2.32   0.026     .0049109     .071665
     firmage |  -.0106424   .0020065    -5.30   0.000    -.0147079   -.0065769
    firmage2 |   .0001163   .0000491     2.37   0.023     .0000168    .0002157
      gender |  -.0110103   .0327599    -0.34   0.739    -.0773881    .0553676
         age |  -.0009863   .0102902    -0.10   0.924    -.0218362    .0198635
        age2 |  -7.93e-06   .0000846    -0.09   0.926    -.0001793    .0001634
      yrinco |   .0040309   .0005348     7.54   0.000     .0029473    .0051146
             |
   education |
          2  |   .0405435   .0485242     0.84   0.409    -.0577759     .138863
          3  |   .0573035   .0334219     1.71   0.095    -.0104157    .1250226
          4  |   .0805295   .0418657     1.92   0.062    -.0042984    .1653574
             |
       _cons |   .9618355   .3255377     2.95   0.005     .3022335    1.621438
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "Full GSS"

added macro:
              e(trust) : "Full GSS"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col3, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trustwvs_sd
(47,109 real changes made, 9,496 to missing)

. eststo col4: /// WVS-based trust
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      15.03
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9014
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0024
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0320269   .0112194     2.85   0.007     .0092942    .0547595
     firmage |  -.0116065   .0020198    -5.75   0.000     -.015699    -.007514
    firmage2 |   .0001147   .0000486     2.36   0.024     .0000163    .0002131
      gender |  -.0142951   .0325372    -0.44   0.663    -.0802217    .0516315
         age |  -.0020615   .0104968    -0.20   0.845      -.02333     .019207
        age2 |   1.75e-06   .0000862     0.02   0.984     -.000173    .0001765
      yrinco |   .0040777   .0005555     7.34   0.000     .0029522    .0052033
             |
   education |
          2  |   .0461817   .0487554     0.95   0.350    -.0526061    .1449696
          3  |   .0616626   .0339439     1.82   0.077    -.0071142    .1304395
          4  |    .089918   .0409993     2.19   0.035     .0068456    .1729904
             |
       _cons |    1.03125   .2971168     3.47   0.001     .4292339    1.633266
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "WVS"

added macro:
              e(trust) : "WVS"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col4, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = trustgps_sd
(37,613 real changes made, 284 to missing)

. eststo col5: /// GPS-based trust
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      16.90
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0389275   .0151411     2.57   0.014     .0082488    .0696063
     firmage |  -.0114843   .0018448    -6.23   0.000    -.0152222   -.0077464
    firmage2 |   .0001165   .0000478     2.43   0.020     .0000195    .0002134
      gender |   -.014942   .0327556    -0.46   0.651    -.0813111    .0514271
         age |  -.0012096   .0102099    -0.12   0.906    -.0218968    .0194777
        age2 |  -4.33e-06   .0000837    -0.05   0.959    -.0001739    .0001653
      yrinco |   .0040323   .0005331     7.56   0.000     .0029521    .0051126
             |
   education |
          2  |   .0456218   .0484806     0.94   0.353    -.0526092    .1438528
          3  |   .0614949   .0344194     1.79   0.082    -.0082454    .1312351
          4  |   .0882536   .0410687     2.15   0.038     .0050406    .1714666
             |
       _cons |    1.10287   .3114155     3.54   0.001     .4718822    1.733858
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "GPS"

added macro:
              e(trust) : "GPS"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col5, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = fair_sd
(37,329 real changes made)

. eststo col6: /// fairness question
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      18.33
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0020
Number of clusters (mainethcode) =         38     Root MSE        =     0.5427

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0654646   .0206409     3.17   0.003     .0236423     .107287
     firmage |  -.0107498   .0019689    -5.46   0.000    -.0147392   -.0067603
    firmage2 |   .0001168    .000049     2.38   0.022     .0000175    .0002162
      gender |  -.0124352   .0329734    -0.38   0.708    -.0792458    .0543753
         age |  -.0014485   .0104595    -0.14   0.891    -.0226414    .0197444
        age2 |  -4.17e-06    .000086    -0.05   0.962    -.0001784      .00017
      yrinco |   .0040532   .0005496     7.37   0.000     .0029396    .0051669
             |
   education |
          2  |   .0421365   .0486489     0.87   0.392    -.0564354    .1407085
          3  |   .0578954   .0334263     1.73   0.092    -.0098326    .1256235
          4  |   .0828974   .0414685     2.00   0.053    -.0011258    .1669205
             |
       _cons |   .8101797   .3041874     2.66   0.011     .1938375    1.426522
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "Fairness"

added macro:
              e(trust) : "Fairness"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col6, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. replace TRUST = helpful_sd
(37,329 real changes made)

. eststo col7: /// helpful intention question
>         reghdfe ash_f1allpat TRUST ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      17.03
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9013
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       TRUST |   .0634803   .0204051     3.11   0.004     .0221357     .104825
     firmage |  -.0108607   .0019766    -5.49   0.000    -.0148658   -.0068556
    firmage2 |   .0001154   .0000496     2.33   0.026     .0000149     .000216
      gender |    -.01396   .0332476    -0.42   0.677     -.081326     .053406
         age |  -.0016049    .010554    -0.15   0.880    -.0229894    .0197795
        age2 |  -2.49e-06   .0000867    -0.03   0.977    -.0001781    .0001731
      yrinco |   .0040694   .0005627     7.23   0.000     .0029292    .0052096
             |
   education |
          2  |   .0460823   .0479302     0.96   0.343    -.0510334    .1431981
          3  |   .0596014   .0330432     1.80   0.079    -.0073505    .1265533
          4  |   .0850715   .0408704     2.08   0.044     .0022602    .1678828
             |
       _cons |    .788309   .2669302     2.95   0.005     .2474571    1.329161
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local trust      "Helpful"

added macro:
              e(trust) : "Helpful"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col7, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. 
. esttab /*using "Table A6_Robustness checks_Panel C.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(18) modelwidth(12) ///
>         mgroups("arsinh(Future patent applications)", pattern(1 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(TRUST) order(TRUST) coeflab(TRUST "CEO's trust") ///
>         stats(patent trust FE controls N nofirms, fmt(%9.3fc %9.0fc %9.0fc) ///
>                 lab("Patent measure" "Trust measure" "Firm \& Year FEs" "Baseline controls" "Observations" "Firms"))

----------------------------------------------------------------------------------------------------------------------------------
                   \multicolumn{7}{c}{arsinh(Future patent applications)}                                                         
                            (1)             (2)             (3)             (4)             (5)             (6)             (7)   
----------------------------------------------------------------------------------------------------------------------------------
CEO's trust               0.055***        0.072***        0.038**         0.032***        0.039**         0.065***        0.063***
                        (0.014)         (0.022)         (0.016)         (0.011)         (0.015)         (0.021)         (0.020)   
----------------------------------------------------------------------------------------------------------------------------------
Patent measure     Granted USPTO                                                                                                   
Trust measure                             LASSO        Full GSS             WVS             GPS        Fairness         Helpful   
Firm \& Year FEs              X               X               X               X               X               X               X   
Baseline controls             X               X               X               X               X               X               X   
Observations             29,384          29,384          29,384          29,384          29,384          29,384          29,384   
Firms                     3,598           3,598           3,598           3,598           3,598           3,598           3,598   
----------------------------------------------------------------------------------------------------------------------------------

. 
. 
. * PANEL D: Alternative sample restrictions
. eststo clear

. eststo col1: /// excluding singletons
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster})
(dropped 189 singleton observations)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,195
Absorbing 2 HDFE groups                           F(  10,     37) =      20.39
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9009
                                                  Adj R-squared   =     0.8877
                                                  Within R-sq.    =     0.0022
Number of clusters (mainethcode) =         38     Root MSE        =     0.5426

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0618858   .0168062     3.68   0.001     .0278332    .0959383
     firmage |  -.0109686   .0018962    -5.78   0.000    -.0148106   -.0071266
    firmage2 |   .0001159   .0000491     2.36   0.024     .0000165    .0002154
      gender |  -.0118746   .0329963    -0.36   0.721    -.0787316    .0549823
         age |  -.0012187   .0102703    -0.12   0.906    -.0220283    .0195909
        age2 |  -6.07e-06   .0000844    -0.07   0.943     -.000177    .0001649
      yrinco |   .0040659   .0005487     7.41   0.000     .0029541    .0051778
             |
   education |
          2  |   .0426374   .0479239     0.89   0.379    -.0544657    .1397406
          3  |   .0592057   .0331928     1.78   0.083    -.0080492    .1264607
          4  |   .0837448   .0410246     2.04   0.048     .0006211    .1668685
             |
       _cons |   .8184672   .2942812     2.78   0.008     .2221968    1.414738
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3409           0        3409     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampexcl   "Singletons"

added macro:
           e(sampexcl) : "Singletons"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29195       3409

.                 eststo col1, add(nofirms r(ndistinct))
(e(nofirms) = 3409 added)

. eststo col2: /// excluding female CEOs
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & gender != 2, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)
note: gender is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)

HDFE Linear regression                            Number of obs   =     28,523
Absorbing 2 HDFE groups                           F(   9,     37) =      23.62
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9008
                                                  Adj R-squared   =     0.8866
                                                  Within R-sq.    =     0.0021
Number of clusters (mainethcode) =         38     Root MSE        =     0.5446

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0580871    .018156     3.20   0.003     .0212995    .0948747
     firmage |  -.0108036   .0017907    -6.03   0.000    -.0144318   -.0071754
    firmage2 |    .000115   .0000525     2.19   0.035     8.57e-06    .0002215
      gender |          0  (omitted)
         age |  -.0031974   .0107071    -0.30   0.767    -.0248921    .0184974
        age2 |     .00001   .0000874     0.11   0.910    -.0001671    .0001871
      yrinco |   .0039424   .0005486     7.19   0.000     .0028309    .0050539
             |
   education |
          2  |   .0419564   .0491137     0.85   0.398    -.0575573    .1414702
          3  |     .05299   .0348617     1.52   0.137    -.0176465    .1236265
          4  |   .0849198   .0438509     1.94   0.060    -.0039306    .1737702
             |
       _cons |   .8883657   .2839785     3.13   0.003     .3129707    1.463761
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3550           0        3550     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampexcl   "Female CEOs"

added macro:
           e(sampexcl) : "Female CEOs"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      28523       3550

.                 eststo col2, add(nofirms r(ndistinct))
(e(nofirms) = 3550 added)

. eststo col3: /// excluding interim CEOs
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & !interimCEO, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     28,909
Absorbing 2 HDFE groups                           F(  10,     37) =      30.13
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9018
                                                  Adj R-squared   =     0.8879
                                                  Within R-sq.    =     0.0026
Number of clusters (mainethcode) =         38     Root MSE        =     0.5411

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0750901   .0189521     3.96   0.000     .0366894    .1134908
     firmage |    -.01147   .0017697    -6.48   0.000    -.0150559   -.0078842
    firmage2 |   .0001188   .0000497     2.39   0.022      .000018    .0002195
      gender |  -.0105502   .0364295    -0.29   0.774    -.0843633    .0632629
         age |  -.0004812   .0096254    -0.05   0.960    -.0199842    .0190217
        age2 |  -.0000154   .0000777    -0.20   0.843    -.0001728    .0001419
      yrinco |   .0043368   .0006527     6.64   0.000     .0030144    .0056593
             |
   education |
          2  |   .0347874   .0493295     0.71   0.485    -.0651636    .1347384
          3  |    .057948   .0350063     1.66   0.106    -.0129816    .1288776
          4  |   .0915327   .0456293     2.01   0.052    -.0009211    .1839864
             |
       _cons |    .742368   .2684263     2.77   0.009     .1984847    1.286251
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3558           0        3558     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampexcl   "Interim CEOs"

added macro:
           e(sampexcl) : "Interim CEOs"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      28909       3558

.                 eststo col3, add(nofirms r(ndistinct))
(e(nofirms) = 3558 added)

. eststo col4: /// 2000-2014
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if year < 2015 & !nonUS, ///
>                 a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     35,956
Absorbing 2 HDFE groups                           F(  10,     37) =      14.91
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8911
                                                  Adj R-squared   =     0.8771
                                                  Within R-sq.    =     0.0020
Number of clusters (mainethcode) =         38     Root MSE        =     0.5650

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0535317   .0150443     3.56   0.001     .0230491    .0840144
     firmage |  -.0020327   .0044187    -0.46   0.648    -.0109858    .0069205
    firmage2 |  -.0000103   .0000311    -0.33   0.742    -.0000733    .0000527
      gender |  -.0297166   .0327196    -0.91   0.370    -.0960129    .0365797
         age |  -.0001908   .0115287    -0.02   0.987    -.0235502    .0231686
        age2 |  -.0000168   .0000967    -0.17   0.863    -.0002128    .0001791
      yrinco |   .0038638    .000516     7.49   0.000     .0028182    .0049094
             |
   education |
          2  |   .0368991   .0484149     0.76   0.451    -.0611988     .134997
          3  |   .0360808   .0349135     1.03   0.308    -.0346607    .1068224
          4  |    .096585   .0406696     2.37   0.023     .0141806    .1789894
             |
       _cons |   .7691401   .3103067     2.48   0.018     .1403991    1.397881
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      4078           0        4078     |
        year |        15           1          14     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampper    "2000-14"

added macro:
            e(sampper) : "2000-14"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      35956       4078

.                 eststo col4, add(nofirms r(ndistinct))
(e(nofirms) = 4078 added)

. eststo col5: /// only patenting firms
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & haspat, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     19,146
Absorbing 2 HDFE groups                           F(  10,     35) =      12.34
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8765
                                                  Adj R-squared   =     0.8597
                                                  Within R-sq.    =     0.0030
Number of clusters (mainethcode) =         36     Root MSE        =     0.6694

                           (Std. err. adjusted for 36 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0802845   .0223225     3.60   0.001     .0349675    .1256015
     firmage |   -.014872   .0048868    -3.04   0.004    -.0247926   -.0049514
    firmage2 |   .0001726   .0000678     2.55   0.015     .0000349    .0003103
      gender |  -.0164895    .053623    -0.31   0.760    -.1253499     .092371
         age |  -.0005425   .0148858    -0.04   0.971    -.0307623    .0296773
        age2 |  -.0000206   .0001215    -0.17   0.867    -.0002673    .0002262
      yrinco |   .0055998   .0008619     6.50   0.000     .0038502    .0073495
             |
   education |
          2  |   .0828108   .0842587     0.98   0.332    -.0882434     .253865
          3  |   .1039315   .0609746     1.70   0.097    -.0198535    .2277165
          4  |   .1371421   .0719478     1.91   0.065    -.0089198     .283204
             |
       _cons |   1.248846   .4284487     2.91   0.006     .3790491    2.118643
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      2266           0        2266     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampincl   "Patent firms"

added macro:
           e(sampincl) : "Patent firms"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      19146       2266

.                 eststo col5, add(nofirms r(ndistinct))
(e(nofirms) = 2266 added)

. eststo col6: /// only during patenting period
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & haspat & inrange(year, firstpatyear-1, lastpatyear), ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     14,881
Absorbing 2 HDFE groups                           F(  10,     34) =       8.90
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8580
                                                  Adj R-squared   =     0.8348
                                                  Within R-sq.    =     0.0040
Number of clusters (mainethcode) =         35     Root MSE        =     0.7313

                           (Std. err. adjusted for 35 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0988474   .0239995     4.12   0.000     .0500747    .1476202
     firmage |  -.0143229   .0058385    -2.45   0.019    -.0261882   -.0024577
    firmage2 |   .0002531   .0000853     2.97   0.005     .0000798    .0004264
      gender |  -.1000973   .0642594    -1.56   0.129    -.2306881    .0304934
         age |  -.0016039    .016422    -0.10   0.923    -.0349775    .0317697
        age2 |  -.0000254   .0001329    -0.19   0.850    -.0002954    .0002447
      yrinco |   .0063288   .0012655     5.00   0.000      .003757    .0089006
             |
   education |
          2  |   .1289577   .1012889     1.27   0.212    -.0768862    .3348016
          3  |   .1458626   .0817586     1.78   0.083     -.020291    .3120161
          4  |   .1921982   .0811367     2.37   0.024     .0273086    .3570878
             |
       _cons |   1.633069   .5264328     3.10   0.004     .5632284    2.702909
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      2068           0        2068     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local sampincl   "Patent period"

added macro:
           e(sampincl) : "Patent period"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      14881       2068

.                 eststo col6, add(nofirms r(ndistinct))
(e(nofirms) = 2068 added)

. eststo col7: /// only patenting years
>         reghdfe log_f1allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin 
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     11,265
Absorbing 2 HDFE groups                           F(  10,     34) =      15.78
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8740
                                                  Adj R-squared   =     0.8468
                                                  Within R-sq.    =     0.0053
Number of clusters (mainethcode) =         35     Root MSE        =     0.6545

                           (Std. err. adjusted for 35 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
log_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0947855   .0268412     3.53   0.001     .0402376    .1493334
     firmage |   .0101782   .0106833     0.95   0.347    -.0115328    .0318891
    firmage2 |   .0001468   .0000975     1.51   0.141    -.0000513    .0003449
      gender |  -.1439822   .0610472    -2.36   0.024    -.2680451   -.0199192
         age |   -.000864   .0165439    -0.05   0.959    -.0344852    .0327572
        age2 |  -.0000296   .0001429    -0.21   0.837    -.0003199    .0002607
      yrinco |   .0074172   .0015428     4.81   0.000     .0042819    .0105524
             |
   education |
          2  |    .133749   .0829358     1.61   0.116    -.0347968    .3022948
          3  |   .1713062   .0819166     2.09   0.044     .0048317    .3377808
          4  |   .1947261   .0748016     2.60   0.014      .042711    .3467412
             |
       _cons |   1.106212    .471061     2.35   0.025     .1489005    2.063523
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      1982           0        1982     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local transform  "ln(.)"

added macro:
          e(transform) : "ln(.)"

.                 estadd local sampincl   "Patent years"

added macro:
           e(sampincl) : "Patent years"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      11265       1982

.                 eststo col7, add(nofirms r(ndistinct))
(e(nofirms) = 1982 added)

. gen D_f1allpat = f1allpat > 0

. eststo col8: /// D_f1allpat as Y var
>         reghdfe D_f1allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =       6.30
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.7154
                                                  Adj R-squared   =     0.6755
                                                  Within R-sq.    =     0.0005
Number of clusters (mainethcode) =         38     Root MSE        =     0.2770

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
  D_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0088248   .0060809     1.45   0.155    -.0034963    .0211459
     firmage |  -.0056675   .0030663    -1.85   0.073    -.0118804    .0005455
    firmage2 |   .0000458   .0000186     2.46   0.019     8.11e-06    .0000835
      gender |   .0161883   .0164965     0.98   0.333    -.0172368    .0496135
         age |   .0012009   .0042953     0.28   0.781    -.0075022     .009904
        age2 |  -.0000153   .0000367    -0.42   0.680    -.0000897    .0000592
      yrinco |   .0004942   .0004496     1.10   0.279    -.0004166    .0014051
             |
   education |
          2  |   .0085998   .0243709     0.35   0.726    -.0407803    .0579799
          3  |   .0142726    .022699     0.63   0.533    -.0317199    .0602651
          4  |   .0209035   .0245909     0.85   0.401    -.0289224    .0707293
             |
       _cons |   .3641125   .1456966     2.50   0.017     .0689031    .6593219
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 sum D_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
  D_f1allpat |     29,384    .3833719    .4862159          0          1

.                 estadd scalar depvarmean = r(mean)

added scalar:
         e(depvarmean) =  .3833719

.                 estadd local transform  "1(.>0)"

added macro:
          e(transform) : "1(.>0)"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col8, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. 
. esttab /*using "Table A6_Robustness checks_Panel D.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(18) modelwidth(12) ///
>         mgroups("Future patent applications", pattern(1 0 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(trust_sd) coeflab(trust_sd "CEO's trust") ///
>         stats(depvarmean transform sampexcl sampper sampincl FE controls N nofirms, fmt(%9.3fc %9.0fc %9.0fc) ///
>                 lab("Dep. var. mean" "Transformation" "Sample excluding" "Sample period" "Sample including" ///
>                         "Firm \& Year FEs" "Baseline controls" "Observations" "Firms"))

--------------------------------------------------------------------------------------------------------------------------------------------------
                   \multicolumn{8}{c}{Future patent applications}                                                                                 
                            (1)             (2)             (3)             (4)             (5)             (6)             (7)             (8)   
--------------------------------------------------------------------------------------------------------------------------------------------------
CEO's trust               0.062***        0.058***        0.075***        0.054***        0.080***        0.099***        0.095***        0.009   
                        (0.017)         (0.018)         (0.019)         (0.015)         (0.022)         (0.024)         (0.027)         (0.006)   
--------------------------------------------------------------------------------------------------------------------------------------------------
Dep. var. mean                                                                                                                            0.383   
Transformation        arsinh(.)       arsinh(.)       arsinh(.)       arsinh(.)       arsinh(.)       arsinh(.)           ln(.)          1(.>0)   
Sample excluding     Singletons     Female CEOs    Interim CEOs                                                                                   
Sample period                                                           2000-14                                                                   
Sample including                                                                   Patent firms    Patent period    Patent years                   
Firm \& Year FEs              X               X               X               X               X               X               X               X   
Baseline controls             X               X               X               X               X               X               X               X   
Observations             29,195          28,523          28,909          35,956          19,146          14,881          11,265          29,384   
Firms                     3,409           3,550           3,558           4,078           2,266           2,068           1,982           3,598   
--------------------------------------------------------------------------------------------------------------------------------------------------

. 
. 
. * PANEL E: Alternative time lags before patent filing   
. eststo clear

. eststo col1: ///
>         reghdfe ash_allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      21.60
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9001
                                                  Adj R-squared   =     0.8861
                                                  Within R-sq.    =     0.0019
Number of clusters (mainethcode) =         38     Root MSE        =     0.5465

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
  ash_allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0685307   .0133865     5.12   0.000      .041407    .0956544
     firmage |  -.0041202   .0059359    -0.69   0.492    -.0161475     .007907
    firmage2 |    .000098   .0000434     2.26   0.030       .00001    .0001861
      gender |   .0144278   .0234464     0.62   0.542    -.0330791    .0619346
         age |  -.0032353   .0087255    -0.37   0.713    -.0209149    .0144442
        age2 |   .0000223   .0000737     0.30   0.764    -.0001271    .0001717
      yrinco |   .0037745    .000701     5.38   0.000     .0023541    .0051949
             |
   education |
          2  |   .0177201   .0471493     0.38   0.709    -.0778134    .1132536
          3  |   .0324117   .0400112     0.81   0.423    -.0486586     .113482
          4  |   .0358086    .045798     0.78   0.439    -.0569869     .128604
             |
       _cons |   .7042296    .216904     3.25   0.002     .2647404    1.143719
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 sum ash_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
ash_f1allpat |     29,384    1.001382    1.617756          0   9.528794

.                 estadd local forward    "0-year"

added macro:
            e(forward) : "0-year"

.                 estadd local sample     "Year T"

added macro:
             e(sample) : "Year T"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col1, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col2: /// 
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-1, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     26,202
Absorbing 2 HDFE groups                           F(  10,     37) =      20.31
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9028
                                                  Adj R-squared   =     0.8875
                                                  Within R-sq.    =     0.0028
Number of clusters (mainethcode) =         38     Root MSE        =     0.5425

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0877082   .0252497     3.47   0.001     .0365475     .138869
     firmage |  -.0144808   .0022546    -6.42   0.000    -.0190491   -.0099125
    firmage2 |   .0001145    .000047     2.44   0.020     .0000193    .0002097
      gender |  -.0104571   .0414169    -0.25   0.802    -.0943757    .0734615
         age |   .0003364   .0118953     0.03   0.978    -.0237658    .0244385
        age2 |   -.000022   .0000983    -0.22   0.824    -.0002212    .0001771
      yrinco |   .0046216   .0007765     5.95   0.000     .0030484    .0061949
             |
   education |
          2  |   .0471808   .0598674     0.79   0.436    -.0741221    .1684838
          3  |   .0770679   .0413567     1.86   0.070    -.0067287    .1608645
          4  |   .1008383   .0528863     1.91   0.064    -.0063196    .2079961
             |
       _cons |   .6973975   .3271285     2.13   0.040     .0345722    1.360223
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3552           0        3552     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local forward    "1-year"

added macro:
            e(forward) : "1-year"

.                 estadd local sample     "T - 1"

added macro:
             e(sample) : "T - 1"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      26202       3552

.                 eststo col2, add(nofirms r(ndistinct))
(e(nofirms) = 3552 added)

. eststo col3: ///
>         reghdfe ash_f2allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      19.27
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8994
                                                  Adj R-squared   =     0.8852
                                                  Within R-sq.    =     0.0031
Number of clusters (mainethcode) =         38     Root MSE        =     0.5473

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f2allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0455151    .024958     1.82   0.076    -.0050547    .0960849
     firmage |  -.0126612   .0063962    -1.98   0.055    -.0256211    .0002988
    firmage2 |   .0001649   .0000571     2.89   0.006     .0000493    .0002806
      gender |  -.0286864   .0245524    -1.17   0.250    -.0784344    .0210615
         age |   .0042374   .0116466     0.36   0.718    -.0193609    .0278357
        age2 |  -.0000562   .0000961    -0.58   0.562     -.000251    .0001386
      yrinco |   .0042472   .0007332     5.79   0.000     .0027616    .0057329
             |
   education |
          2  |  -.0069016   .0419733    -0.16   0.870    -.0919475    .0781443
          3  |   .0064978   .0339929     0.19   0.849    -.0623784     .075374
          4  |   .0794375   .0346234     2.29   0.028      .009284    .1495911
             |
       _cons |   .7915854   .3007566     2.63   0.012     .1821946    1.400976
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local forward    "2-year"

added macro:
            e(forward) : "2-year"

.                 estadd local sample     "T"

added macro:
             e(sample) : "T"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col3, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col4: ///
>         reghdfe ash_f2allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-1, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     26,202
Absorbing 2 HDFE groups                           F(  10,     37) =      19.90
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9005
                                                  Adj R-squared   =     0.8849
                                                  Within R-sq.    =     0.0044
Number of clusters (mainethcode) =         38     Root MSE        =     0.5488

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f2allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |     .07558   .0266826     2.83   0.007      .021516     .129644
     firmage |  -.0218064   .0075993    -2.87   0.007     -.037204   -.0064088
    firmage2 |   .0001626    .000059     2.75   0.009      .000043    .0002822
      gender |   -.044613   .0362229    -1.23   0.226    -.1180076    .0287816
         age |   .0018235   .0125021     0.15   0.885    -.0235081    .0271551
        age2 |  -.0000417   .0001017    -0.41   0.684    -.0002477    .0001642
      yrinco |   .0053308   .0008057     6.62   0.000     .0036984    .0069633
             |
   education |
          2  |   .0209363    .061508     0.34   0.735    -.1036906    .1455633
          3  |   .0438093   .0511669     0.86   0.397    -.0598648    .1474833
          4  |   .1389006   .0517485     2.68   0.011     .0340481    .2437531
             |
       _cons |   .8713651   .3500455     2.49   0.017     .1621056    1.580625
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3552           0        3552     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local forward    "2-year"

added macro:
            e(forward) : "2-year"

.                 estadd local sample     "T-1"

added macro:
             e(sample) : "T-1"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      26202       3552

.                 eststo col4, add(nofirms r(ndistinct))
(e(nofirms) = 3552 added)

. eststo col5: ///
>         reghdfe ash_f2allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-2, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     20,793
Absorbing 2 HDFE groups                           F(  10,     37) =      24.17
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9116
                                                  Adj R-squared   =     0.8948
                                                  Within R-sq.    =     0.0033
Number of clusters (mainethcode) =         38     Root MSE        =     0.5275

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f2allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0884244   .0359672     2.46   0.019      .015548    .1613008
     firmage |  -.0129985   .0023608    -5.51   0.000    -.0177818   -.0082151
    firmage2 |   .0001124   .0000692     1.62   0.113    -.0000279    .0002528
      gender |  -.0562912   .0527617    -1.07   0.293    -.1631967    .0506142
         age |   .0042728   .0157082     0.27   0.787     -.027555    .0361007
        age2 |  -.0000634   .0001283    -0.49   0.624    -.0003234    .0001965
      yrinco |   .0056884    .000891     6.38   0.000     .0038831    .0074936
             |
   education |
          2  |   .0482339   .0743817     0.65   0.521    -.1024777    .1989454
          3  |   .0717172   .0632543     1.13   0.264    -.0564482    .1998827
          4  |   .1444896   .0645183     2.24   0.031     .0137632    .2752161
             |
       _cons |   .5992581   .4421356     1.36   0.184    -.2965937     1.49511
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3301           0        3301     |
        year |        11           1          10     |
-----------------------------------------------------+

.                 estadd local forward    "2-year"

added macro:
            e(forward) : "2-year"

.                 estadd local sample     "T-2"

added macro:
             e(sample) : "T-2"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      20793       3301

.                 eststo col5, add(nofirms r(ndistinct))
(e(nofirms) = 3301 added)

. eststo col6: ///
>         reghdfe ash_f3allpat trust_sd ${firm} ${ceo} if ${sample}, ///
>         a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     29,384
Absorbing 2 HDFE groups                           F(  10,     37) =      14.46
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8966
                                                  Adj R-squared   =     0.8820
                                                  Within R-sq.    =     0.0032
Number of clusters (mainethcode) =         38     Root MSE        =     0.5528

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f3allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0387184   .0221166     1.75   0.088    -.0060941    .0835309
     firmage |  -.0213676   .0073281    -2.92   0.006    -.0362158   -.0065194
    firmage2 |    .000184   .0000598     3.08   0.004     .0000629    .0003051
      gender |  -.0198088   .0200546    -0.99   0.330    -.0604434    .0208257
         age |   .0048031   .0121674     0.39   0.695    -.0198503    .0294566
        age2 |  -.0000598   .0001044    -0.57   0.570    -.0002714    .0001517
      yrinco |   .0035797   .0010306     3.47   0.001     .0014915    .0056678
             |
   education |
          2  |  -.0481046   .0403129    -1.19   0.240    -.1297862     .033577
          3  |  -.0272467   .0365383    -0.75   0.461    -.1012802    .0467869
          4  |   .0514267   .0471691     1.09   0.283     -.044147    .1470003
             |
       _cons |   .9682109   .2717926     3.56   0.001     .4175067    1.518915
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3598           0        3598     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local forward    "3-year"

added macro:
            e(forward) : "3-year"

.                 estadd local sample     "T"

added macro:
             e(sample) : "T"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      29384       3598

.                 eststo col6, add(nofirms r(ndistinct))
(e(nofirms) = 3598 added)

. eststo col7: ///
>         reghdfe ash_f3allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-1, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     26,202
Absorbing 2 HDFE groups                           F(  10,     37) =      15.90
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8978
                                                  Adj R-squared   =     0.8817
                                                  Within R-sq.    =     0.0038
Number of clusters (mainethcode) =         38     Root MSE        =     0.5538

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f3allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0476847   .0272274     1.75   0.088    -.0074832    .1028526
     firmage |  -.0197006   .0071523    -2.75   0.009    -.0341926   -.0052087
    firmage2 |     .00019    .000065     2.92   0.006     .0000584    .0003216
      gender |  -.0470753   .0257356    -1.83   0.075    -.0992205    .0050699
         age |    .005973   .0137281     0.44   0.666    -.0218427    .0337888
        age2 |  -.0000709    .000117    -0.61   0.548    -.0003081    .0001662
      yrinco |   .0038906   .0010629     3.66   0.001      .001737    .0060441
             |
   education |
          2  |  -.0521169   .0445577    -1.17   0.250    -.1423994    .0381657
          3  |  -.0272055   .0419804    -0.65   0.521    -.1122659    .0578549
          4  |   .0789134   .0531569     1.48   0.146    -.0287927    .1866195
             |
       _cons |   .8802062   .3172411     2.77   0.009     .2374146    1.522998
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3552           0        3552     |
        year |        12           1          11     |
-----------------------------------------------------+

.                 estadd local forward    "3-year"

added macro:
            e(forward) : "3-year"

.                 estadd local sample     "T-1"

added macro:
             e(sample) : "T-1"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      26202       3552

.                 eststo col7, add(nofirms r(ndistinct))
(e(nofirms) = 3552 added)

. eststo col8: ///
>         reghdfe ash_f3allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-2, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     20,793
Absorbing 2 HDFE groups                           F(  10,     37) =      19.26
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9094
                                                  Adj R-squared   =     0.8922
                                                  Within R-sq.    =     0.0039
Number of clusters (mainethcode) =         38     Root MSE        =     0.5339

                           (Std. err. adjusted for 38 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f3allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0859106   .0278182     3.09   0.004     .0295456    .1422757
     firmage |  -.0140571   .0020673    -6.80   0.000    -.0182459   -.0098682
    firmage2 |   .0001618   .0000854     1.89   0.066    -.0000113    .0003349
      gender |   -.113463   .0418768    -2.71   0.010    -.1983135   -.0286125
         age |    .008091   .0162608     0.50   0.622    -.0248565    .0410385
        age2 |  -.0000929   .0001373    -0.68   0.503     -.000371    .0001852
      yrinco |    .005202   .0013337     3.90   0.000     .0024996    .0079044
             |
   education |
          2  |  -.0310413   .0611896    -0.51   0.615    -.1550231    .0929406
          3  |  -.0007426    .055553    -0.01   0.989    -.1133036    .1118185
          4  |   .0927489   .0506195     1.83   0.075    -.0098159    .1953137
             |
       _cons |   .5983243   .4379858     1.37   0.180    -.2891192    1.485768
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3301           0        3301     |
        year |        11           1          10     |
-----------------------------------------------------+

.                 estadd local forward    "3-year"

added macro:
            e(forward) : "3-year"

.                 estadd local sample     "T-2"

added macro:
             e(sample) : "T-2"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      20793       3301

.                 eststo col8, add(nofirms r(ndistinct))
(e(nofirms) = 3301 added)

. eststo col9: ///
>         reghdfe ash_f3allpat trust_sd ${firm} ${ceo} if ${sample} ///
>         & year <= termendyr-3, a(boardid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     16,242
Absorbing 2 HDFE groups                           F(  10,     36) =      11.45
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9219
                                                  Adj R-squared   =     0.9040
                                                  Within R-sq.    =     0.0035
Number of clusters (mainethcode) =         37     Root MSE        =     0.5074

                           (Std. err. adjusted for 37 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f3allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .1342643   .0481201     2.79   0.008     .0366723    .2318563
     firmage |  -.0134362    .002556    -5.26   0.000    -.0186201   -.0082523
    firmage2 |   .0001189   .0000959     1.24   0.223    -.0000756    .0003134
      gender |  -.1677881   .0562982    -2.98   0.005    -.2819662   -.0536101
         age |    .011457   .0186846     0.61   0.544     -.026437    .0493511
        age2 |  -.0001278   .0001547    -0.83   0.414    -.0004414    .0001859
      yrinco |   .0055398   .0011546     4.80   0.000     .0031982    .0078814
             |
   education |
          2  |   .0010468   .0886161     0.01   0.991    -.1786751    .1807686
          3  |   .0428153   .0895118     0.48   0.635    -.1387231    .2243537
          4  |   .0880698    .074508     1.18   0.245    -.0630395     .239179
             |
       _cons |   .3298891   .4904286     0.67   0.505    -.6647462    1.324524
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     boardid |      3013           0        3013     |
        year |        10           1           9     |
-----------------------------------------------------+

.                 estadd local forward    "3-year"

added macro:
            e(forward) : "3-year"

.                 estadd local sample     "T-3"

added macro:
             e(sample) : "T-3"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct boardid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 boardid |      16242       3013

.                 eststo col9, add(nofirms r(ndistinct))
(e(nofirms) = 3013 added)

. 
. esttab /*using "Table A6_Robustness checks_Panel E.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(18) modelwidth(12) ///
>         mgroups("arsinh(Future patent applications)", pattern(1 0 0 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(trust_sd) coeflab(trust_sd "CEO's trust") ///
>         stats(forward sample FE controls N nofirms, fmt(%9.0fc %9.0fc) ///
>                 lab("Forward" "Sample including" "Firm \& Year FEs" "Baseline controls" "Observations" "Firms"))

------------------------------------------------------------------------------------------------------------------------------------------------------------------
                   \multicolumn{9}{c}{arsinh(Future patent applications)}                                                                                         
                            (1)             (2)             (3)             (4)             (5)             (6)             (7)             (8)             (9)   
------------------------------------------------------------------------------------------------------------------------------------------------------------------
CEO's trust               0.069***        0.088***        0.046*          0.076***        0.088**         0.039*          0.048*          0.086***        0.134***
                        (0.013)         (0.025)         (0.025)         (0.027)         (0.036)         (0.022)         (0.027)         (0.028)         (0.048)   
------------------------------------------------------------------------------------------------------------------------------------------------------------------
Forward                  0-year          1-year          2-year          2-year          2-year          3-year          3-year          3-year          3-year   
Sample including         Year T           T - 1               T             T-1             T-2               T             T-1             T-2             T-3   
Firm \& Year FEs              X               X               X               X               X               X               X               X               X   
Baseline controls             X               X               X               X               X               X               X               X               X   
Observations             29,384          26,202          29,384          26,202          20,793          29,384          26,202          20,793          16,242   
Firms                     3,598           3,552           3,598           3,552           3,301           3,598           3,552           3,301           3,013   
------------------------------------------------------------------------------------------------------------------------------------------------------------------

. 
.         
. * PANEL F: Alternative CEO transition event sample
. use "CEO transition event sample.dta", clear

. *** construct additional variables
. replace moretrust = 9 if moretrust + lesstrust == 0
(1,166 real changes made)

. gen D_f1allpat = f1allpat > 0

. gen postxdeltatrust = postchange * deltatrust if trust_sd != .
(111 missing values generated)

. gen postxtrustbf = postchange * trustbf if trust_sd != .
(111 missing values generated)

. *** set globals
. global firm = "firmage firmage2"

. global ceo = "gender age age2 yrinco i.education"

. global eventsample = "!bothnonUS"

. global cluster = "mainethcode"

. 
. eststo clear

. eststo col1: /// baseline
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${eventsample}, ///
>         a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     20,389
Absorbing 2 HDFE groups                           F(  10,     36) =      52.10
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9084
                                                  Adj R-squared   =     0.8958
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         37     Root MSE        =     0.5499

                           (Std. err. adjusted for 37 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0670031   .0176954     3.79   0.001     .0311152    .1028909
     firmage |  -.0104239   .0015468    -6.74   0.000     -.013561   -.0072868
    firmage2 |   .0000425    .000085     0.50   0.620    -.0001298    .0002148
      gender |  -.0319264   .0291859    -1.09   0.281    -.0911181    .0272653
         age |   .0007568   .0062304     0.12   0.904    -.0118791    .0133926
        age2 |  -.0000272   .0000539    -0.50   0.618    -.0001366    .0000822
      yrinco |   .0028729   .0006123     4.69   0.000     .0016311    .0041146
             |
   education |
          2  |   .0857167    .040717     2.11   0.042     .0031388    .1682947
          3  |   .0928836   .0369665     2.51   0.017     .0179121    .1678551
          4  |   .1325649   .0285102     4.65   0.000     .0747435    .1903863
             |
       _cons |   .9084103   .1605669     5.66   0.000     .5827655    1.234055
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2446           0        2446     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      20389       2446

.                 eststo col1, add(noevents r(ndistinct))
(e(noevents) = 2446 added)

. eststo col2: /// by change in trust
>         reghdfe ash_f1allpat c.trust_sd##c.deltatrust ${firm} ${ceo} if ${eventsample}, ///
>         a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)
note: deltatrust is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)

HDFE Linear regression                            Number of obs   =     20,389
Absorbing 2 HDFE groups                           F(  11,     36) =      47.24
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9084
                                                  Adj R-squared   =     0.8958
                                                  Within R-sq.    =     0.0023
Number of clusters (mainethcode) =         37     Root MSE        =     0.5499

                                      (Std. err. adjusted for 37 clusters in mainethcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
------------------------+----------------------------------------------------------------
               trust_sd |   .0673432   .0177997     3.78   0.001     .0312437    .1034427
             deltatrust |          0  (omitted)
                        |
c.trust_sd#c.deltatrust |   .0047687   .0108858     0.44   0.664    -.0173088    .0268461
                        |
                firmage |  -.0104256   .0015489    -6.73   0.000    -.0135669   -.0072843
               firmage2 |   .0000427    .000085     0.50   0.619    -.0001297    .0002151
                 gender |  -.0322023    .029333    -1.10   0.280    -.0916925    .0272878
                    age |   .0007533   .0062285     0.12   0.904    -.0118787    .0133853
                   age2 |  -.0000271   .0000539    -0.50   0.619    -.0001364    .0000823
                 yrinco |   .0029178   .0005943     4.91   0.000     .0017126     .004123
                        |
              education |
                     2  |   .0860808   .0406697     2.12   0.041     .0035989    .1685627
                     3  |   .0932185   .0368821     2.53   0.016     .0184181    .1680189
                     4  |   .1328768   .0284694     4.67   0.000     .0751381    .1906154
                        |
                  _cons |    .906081    .159535     5.68   0.000      .582529    1.229633
-----------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2446           0        2446     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      20389       2446

.                 eststo col2, add(noevents r(ndistinct))
(e(noevents) = 2446 added)

. eststo col3: /// by trust-decreasing vs. trust-increasing
>         reghdfe ash_f1allpat c.trust_sd#i.moretrust ${firm} ${ceo} if ${eventsample}, ///
>         a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)
note: 9bn.moretrust#c.trust_sd is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)

HDFE Linear regression                            Number of obs   =     20,389
Absorbing 2 HDFE groups                           F(  11,     36) =      48.96
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9084
                                                  Adj R-squared   =     0.8958
                                                  Within R-sq.    =     0.0024
Number of clusters (mainethcode) =         37     Root MSE        =     0.5499

                                   (Std. err. adjusted for 37 clusters in mainethcode)
--------------------------------------------------------------------------------------
                     |               Robust
        ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
---------------------+----------------------------------------------------------------
moretrust#c.trust_sd |
                  0  |   .0775298   .0200259     3.87   0.000     .0369153    .1181443
                  1  |   .0553057   .0276127     2.00   0.053    -.0006956    .1113069
                  9  |          0  (omitted)
                     |
             firmage |  -.0104564   .0015266    -6.85   0.000    -.0135524   -.0073603
            firmage2 |   .0000415    .000085     0.49   0.628    -.0001309    .0002138
              gender |  -.0312684   .0295913    -1.06   0.298    -.0912824    .0287456
                 age |   .0006672   .0062022     0.11   0.915    -.0119115    .0132459
                age2 |  -.0000267   .0000539    -0.50   0.623     -.000136    .0000825
              yrinco |   .0027654   .0006004     4.61   0.000     .0015477    .0039831
                     |
           education |
                  2  |   .0850464    .040233     2.11   0.042       .00345    .1666428
                  3  |   .0922646   .0364789     2.53   0.016      .018282    .1662472
                  4  |   .1316748   .0280893     4.69   0.000     .0747071    .1886426
                     |
               _cons |   .9321149   .1577338     5.91   0.000      .612216    1.252014
--------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2446           0        2446     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      20389       2446

.                 eststo col3, add(noevents r(ndistinct))
(e(noevents) = 2446 added)

. eststo col4: /// non-patenting pre-transition
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${eventsample} ///
>         & haspatpre == 0, a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =      8,788
Absorbing 2 HDFE groups                           F(  10,     31) =       4.37
Statistics robust to heteroskedasticity           Prob > F        =     0.0007
                                                  R-squared       =     0.4156
                                                  Adj R-squared   =     0.3275
                                                  Within R-sq.    =     0.0030
Number of clusters (mainethcode) =         32     Root MSE        =     0.2152

                           (Std. err. adjusted for 32 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |  -.0022039   .0128258    -0.17   0.865    -.0283622    .0239544
     firmage |   -.000843   .0041967    -0.20   0.842    -.0094022    .0077162
    firmage2 |   -.000078   .0000375    -2.08   0.046    -.0001546   -1.46e-06
      gender |   .0274331   .0266302     1.03   0.311    -.0268795    .0817456
         age |    .000264   .0045045     0.06   0.954    -.0089229    .0094509
        age2 |  -.0000114   .0000391    -0.29   0.772    -.0000911    .0000683
      yrinco |   .0000299   .0003646     0.08   0.935    -.0007137    .0007734
             |
   education |
          2  |   .0023279   .0138684     0.17   0.868    -.0259569    .0306128
          3  |  -.0160018   .0124616    -1.28   0.209    -.0414175    .0094138
          4  |  -.0274879    .013794    -1.99   0.055    -.0556209    .0006451
             |
       _cons |    .133618   .1233024     1.08   0.287     -.117859    .3850949
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      1130           0        1130     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local haspatpre  "No"

added macro:
          e(haspatpre) : "No"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |       8788       1130

.                 eststo col4, add(noevents r(ndistinct))
(e(noevents) = 1130 added)

. eststo col5: /// patenting pre-transition
>         reghdfe ash_f1allpat trust_sd ${firm} ${ceo} if ${eventsample} ///
>         & haspatpre == 1, a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     11,601
Absorbing 2 HDFE groups                           F(  10,     34) =      14.43
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8789
                                                  Adj R-squared   =     0.8631
                                                  Within R-sq.    =     0.0043
Number of clusters (mainethcode) =         35     Root MSE        =     0.6973

                           (Std. err. adjusted for 35 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .1062836   .0222952     4.77   0.000     .0609742     .151593
     firmage |   -.005725   .0038763    -1.48   0.149    -.0136025    .0021526
    firmage2 |   .0001205    .000128     0.94   0.353    -.0001397    .0003806
      gender |   -.072889   .0454452    -1.60   0.118    -.1652448    .0194668
         age |   .0033521   .0106087     0.32   0.754    -.0182073    .0249115
        age2 |  -.0000561    .000096    -0.58   0.563    -.0002512     .000139
      yrinco |   .0042334   .0010185     4.16   0.000     .0021635    .0063033
             |
   education |
          2  |   .1865032   .0601551     3.10   0.004     .0642533    .3087532
          3  |   .2021139   .0557855     3.62   0.001     .0887441    .3154837
          4  |   .2794423   .0443642     6.30   0.000     .1892835    .3696011
             |
       _cons |   1.212746   .3114599     3.89   0.000     .5797836    1.845709
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      1316           0        1316     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local haspatpre  "Yes"

added macro:
          e(haspatpre) : "Yes"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      11601       1316

.                 eststo col5, add(noevents r(ndistinct))
(e(noevents) = 1316 added)

. eststo col6: /// extensive margin among patenting pre-transition
>         reghdfe D_f1allpat trust_sd ${firm} ${ceo} if ${eventsample} ///
>         & haspatpre == 1, a(eventid year) cluster(${cluster}) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     11,601
Absorbing 2 HDFE groups                           F(  10,     34) =       6.58
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.5495
                                                  Adj R-squared   =     0.4908
                                                  Within R-sq.    =     0.0020
Number of clusters (mainethcode) =         35     Root MSE        =     0.3307

                           (Std. err. adjusted for 35 clusters in mainethcode)
------------------------------------------------------------------------------
             |               Robust
  D_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    trust_sd |   .0212175   .0103949     2.04   0.049     .0000925    .0423424
     firmage |  -.0122757   .0060858    -2.02   0.052    -.0246435    .0000921
    firmage2 |   .0000911   .0000263     3.46   0.001     .0000376    .0001445
      gender |   .0159748   .0237531     0.67   0.506    -.0322973    .0642469
         age |   .0030343   .0037759     0.80   0.427    -.0046392    .0107078
        age2 |  -.0000255   .0000337    -0.76   0.455     -.000094     .000043
      yrinco |   .0004642   .0004153     1.12   0.271    -.0003797    .0013081
             |
   education |
          2  |   .0050259   .0294839     0.17   0.866    -.0548926    .0649444
          3  |   .0125802   .0334256     0.38   0.709    -.0553488    .0805092
          4  |   .0447982   .0279699     1.60   0.118    -.0120436    .1016399
             |
       _cons |   .6516773   .1624066     4.01   0.000     .3216274    .9817272
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      1316           0        1316     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 sum D_f1allpat if e(sample)

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
  D_f1allpat |     11,601    .6876993    .4634517          0          1

.                 estadd scalar depvarmean = r(mean)

added scalar:
         e(depvarmean) =  .68769934

.                 estadd local transform  "1(.>0)"

added macro:
          e(transform) : "1(.>0)"

.                 estadd local haspatpre  "Yes"

added macro:
          e(haspatpre) : "Yes"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      11601       1316

.                 eststo col6, add(noevents r(ndistinct))
(e(noevents) = 1316 added)

. eststo col7: /// preceding CEO's trust as IV for change in trust, RF
>         reghdfe ash_f1allpat postxtrustbf postchange ${firm} ${ceo} if ${eventsample}, ///
>         a(eventid year) cluster(mainethcoder) keepsin
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     20,389
Absorbing 2 HDFE groups                           F(  11,     67) =      14.58
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.9084
                                                  Adj R-squared   =     0.8958
                                                  Within R-sq.    =     0.0022
Number of clusters (mainethcoder) =         68    Root MSE        =     0.5500

                          (Std. err. adjusted for 68 clusters in mainethcoder)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
postxtrustbf |  -.0668039    .023076    -2.89   0.005    -.1128637    -.020744
  postchange |   .3060901   .1227012     2.49   0.015     .0611775    .5510026
     firmage |  -.0100785   .0025661    -3.93   0.000    -.0152004   -.0049566
    firmage2 |   .0000371   .0000759     0.49   0.627    -.0001145    .0001887
      gender |  -.0307291   .0349675    -0.88   0.383    -.1005246    .0390664
         age |    .000252   .0058602     0.04   0.966     -.011445     .011949
        age2 |  -.0000274   .0000581    -0.47   0.638    -.0001434    .0000885
      yrinco |    .002168   .0010376     2.09   0.040     .0000969    .0042392
             |
   education |
          2  |   .0864217   .0447917     1.93   0.058    -.0029829    .1758263
          3  |   .0916531    .040072     2.29   0.025     .0116691    .1716372
          4  |   .1301804   .0438464     2.97   0.004     .0426626    .2176982
             |
       _cons |   1.297476   .1938248     6.69   0.000        .9106    1.684352
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2446           0        2446     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local spec       "RF"

added macro:
               e(spec) : "RF"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      20389       2446

.                 eststo col7, add(noevents r(ndistinct))
(e(noevents) = 2446 added)

. eststo col8: /// preceding CEO's trust as IV for change in trust, IV
>         ivreghdfe ash_f1allpat (postxdeltatrust = postxtrustbf) postchange ${firm} ${ceo} ///
>         if ${eventsample}, a(eventid year) cluster(mainethcoder) keepsin first
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 7 iterations)

First-stage regressions
-----------------------


First-stage regression of postxdeltatrust:

Statistics robust to heteroskedasticity and clustering on mainethcoder
Number of obs =                  20389
Number of clusters (mainethcoder) =     68
------------------------------------------------------------------------------
             |               Robust
postxdelta~t | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
postxtrustbf |  -.9890308   .0192983   -51.25   0.000    -1.026857   -.9512042
  postchange |   5.017595   .0994198    50.47   0.000     4.822723    5.212467
     firmage |    .004971   .0020719     2.40   0.016     .0009099    .0090322
    firmage2 |   .0000557   .0000226     2.47   0.014     .0000115       .0001
      gender |  -.0188801   .0257512    -0.73   0.463     -.069355    .0315948
         age |  -.0004624   .0055771    -0.08   0.934    -.0113941    .0104692
        age2 |   8.09e-06   .0000537     0.15   0.880    -.0000971    .0001133
      yrinco |   .0000442    .000514     0.09   0.932    -.0009633    .0010516
             |
   education |
          0  |          0  (empty)
          2  |   .0012317   .0322332     0.04   0.970    -.0619485    .0644118
          3  |  -.0207441   .0289184    -0.72   0.473    -.0774269    .0359387
          4  |  -.0148501   .0315608    -0.47   0.638    -.0767124    .0470121
------------------------------------------------------------------------------
F test of excluded instruments:
  F(  1,    67) =  2626.51
  Prob > F      =   0.0000
Sanderson-Windmeijer multivariate F test of excluded instruments:
  F(  1,    67) =  2626.51
  Prob > F      =   0.0000



Summary results for first-stage regressions
-------------------------------------------

                                           (Underid)            (Weak id)
Variable     | F(  1,    67)  P-val | SW Chi-sq(  1) P-val | SW F(  1,    67)
postxdeltatr |    2626.51    0.0000 |     3032.85   0.0000 |     2626.51

NB: first-stage test statistics cluster-robust

Stock-Yogo weak ID F test critical values for single endogenous regressor:
                                   10% maximal IV size             16.38
                                   15% maximal IV size              8.96
                                   20% maximal IV size              6.66
                                   25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for i.i.d. errors only.

Underidentification test
Ho: matrix of reduced form coefficients has rank=K1-1 (underidentified)
Ha: matrix has rank=K1 (identified)
Kleibergen-Paap rk LM statistic          Chi-sq(1)=15.24    P-val=0.0001

Weak identification test
Ho: equation is weakly identified
Cragg-Donald Wald F statistic                                   14690.32
Kleibergen-Paap Wald rk F statistic                              2626.51

Stock-Yogo weak ID test critical values for K1=1 and L1=1:
                                   10% maximal IV size             16.38
                                   15% maximal IV size              8.96
                                   20% maximal IV size              6.66
                                   25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.

Weak-instrument-robust inference
Tests of joint significance of endogenous regressors B1 in main equation
Ho: B1=0 and orthogonality conditions are valid
Anderson-Rubin Wald test           F(1,67)=        8.38     P-val=0.0051
Anderson-Rubin Wald test           Chi-sq(1)=      9.68     P-val=0.0019
Stock-Wright LM S statistic        Chi-sq(1)=      8.12     P-val=0.0044

NB: Underidentification, weak identification and weak-identification-robust
    test statistics cluster-robust

Number of clusters             N_clust  =         68
Number of observations               N  =      20389
Number of regressors                 K  =         11
Number of endogenous regressors      K1 =          1
Number of instruments                L  =         11
Number of excluded instruments       L1 =          1

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on mainethcoder

Number of clusters (mainethcoder) =     68            Number of obs =    20389
                                                      F( 11,    67) =    14.28
                                                      Prob > F      =   0.0000
Total (centered) SS     =  5432.458843                Centered R2   =   0.0026
Total (uncentered) SS   =  5432.458843                Uncentered R2 =   0.0026
Residual SS             =  5418.604345                Root MSE      =    .5499

---------------------------------------------------------------------------------
                |               Robust
   ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
----------------+----------------------------------------------------------------
postxdeltatrust |   .0675448    .023224     2.91   0.005     .0211895       .1139
     postchange |  -.0328223   .0132798    -2.47   0.016    -.0593289   -.0063156
        firmage |  -.0104143   .0025239    -4.13   0.000    -.0154519   -.0053766
       firmage2 |   .0000333   .0000759     0.44   0.662    -.0001181    .0001847
         gender |  -.0294538   .0352922    -0.83   0.407    -.0998974    .0409897
            age |   .0002832   .0056848     0.05   0.960    -.0110637    .0116302
           age2 |   -.000028    .000056    -0.50   0.619    -.0001397    .0000837
         yrinco |   .0021651   .0010356     2.09   0.040     .0000981    .0042321
                |
      education |
             0  |          0  (empty)
             2  |   .0863385   .0461761     1.87   0.066    -.0058294    .1785063
             3  |   .0930543   .0412978     2.25   0.028     .0106235    .1754851
             4  |   .1311835    .044987     2.92   0.005     .0413891    .2209778
---------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):             15.236
                                                   Chi-sq(1) P-val =    0.0001
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):              1.5e+04
                         (Kleibergen-Paap rk Wald F statistic):       2626.513
Stock-Yogo weak ID test critical values: 10% maximal IV size             16.38
                                         15% maximal IV size              8.96
                                         20% maximal IV size              6.66
                                         25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         postxdeltatrust
Included instruments: postchange firmage firmage2 gender age age2 yrinco
                      2.education 3.education 4.education
Excluded instruments: postxtrustbf
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2446           0        2446     |
        year |        13           1          12     |
-----------------------------------------------------+

.                 estadd local transform  "arsinh(.)"

added macro:
          e(transform) : "arsinh(.)"

.                 estadd local spec       "IV"

added macro:
               e(spec) : "IV"

.                 estadd local FE         "X"

added macro:
                 e(FE) : "X"

.                 estadd local controls   "X"

added macro:
           e(controls) : "X"

.                 distinct eventid if e(sample)

         |        Observations
         |      total   distinct
---------+----------------------
 eventid |      20389       2446

.                 eststo col8, add(noevents r(ndistinct))
(e(noevents) = 2446 added)

. 
. global coeflist1 = "trust_sd c.trust_sd#c.deltatrust 0.moretrust#c.trust_sd 1.moretrust#c.trust_sd"

. global coeflist2 = "postxtrustbf postxdeltatrust"

. esttab /*using "Table A6_Robustness checks_Panel F.tex"*/, ///
>         cells(b(star fmt(3)) se(par fmt(3))) starlevels(* 0.1 ** 0.05 *** 0.01) ///
>         nomtitle collabels(none) label varwidth(35) modelwidth(12) ///
>         mgroups("Future patent applications", pattern(1 0 0 0 0 0 0 0) ///
>                 prefix(\multicolumn{@span}{c}{) suffix(}) span) ///
>         keep(${coeflist1} ${coeflist2}) order(${coeflist1} ${coeflist2}) ///
>         coeflab(trust_sd "CEO's trust" ///
>                         c.trust_sd#c.deltatrust "Trust $\times$ Change in trust" ///
>                         0.moretrust#c.trust_sd "Trust $\times$ 1(Trust-decreasing)" ///
>                         1.moretrust#c.trust_sd "Trust $\times$ 1(Trust-increasing)" ///
>                         postxtrustbf "Post $\times$ Preceding CEO's trust" ///
>                         postxdeltatrust "Post $\times$ Change in trust") ///
>         stats(depvarmean transform spec haspatpre FE controls N noevents, fmt(%9.3fc %9.0fc %9.0fc) ///
>                 lab("Dep. var. mean" "Tranformation" "Specificaiton" "Patenting pre-transition" ///
>                         "Event \& Year FEs" "Baseline controls" "Observations" "Events"))

-------------------------------------------------------------------------------------------------------------------------------------------------------------------
                                    \multicolumn{8}{c}{Future patent applications}                                                                                 
                                             (1)             (2)             (3)             (4)             (5)             (6)             (7)             (8)   
-------------------------------------------------------------------------------------------------------------------------------------------------------------------
CEO's trust                                0.067***        0.067***                       -0.002           0.106***        0.021**                                 
                                         (0.018)         (0.018)                         (0.013)         (0.022)         (0.010)                                   
Trust $\times$ Change in trust                             0.005                                                                                                   
                                                         (0.011)                                                                                                   
Trust $\times$ 1(Trust-decreasing)                                         0.078***                                                                                
                                                                         (0.020)                                                                                   
Trust $\times$ 1(Trust-increasing)                                         0.055*                                                                                  
                                                                         (0.028)                                                                                   
Post $\times$ Preceding CEO's trust                                                                                                       -0.067***                
                                                                                                                                         (0.023)                   
Post $\times$ Change in trust                                                                                                                              0.068***
                                                                                                                                                         (0.023)   
-------------------------------------------------------------------------------------------------------------------------------------------------------------------
Dep. var. mean                                                                                                             0.688                                   
Tranformation                          arsinh(.)       arsinh(.)       arsinh(.)       arsinh(.)       arsinh(.)          1(.>0)       arsinh(.)       arsinh(.)   
Specificaiton                                                                                                                                 RF              IV   
Patenting pre-transition                                                                      No             Yes             Yes                                   
Event \& Year FEs                              X               X               X               X               X               X               X               X   
Baseline controls                              X               X               X               X               X               X               X               X   
Observations                              20,389          20,389          20,389           8,788          11,601          11,601          20,389          20,389   
Events                                     2,446           2,446           2,446           1,130           1,316           1,316           2,446           2,446   
-------------------------------------------------------------------------------------------------------------------------------------------------------------------

. 
. 
. * PANEL G: Alternative difference-in-differences specifications
. use "CEO transition event sample.dta", clear

. keep if !bothnonUS & trust_sd != .
(111 observations deleted)

. *** construct additional variables
. gen tempyear = year if postchange == 1
(12,005 missing values generated)

. bysort eventid: egen cohort = min(tempyear)
(141 missing values generated)

. drop tempyear

. *** clean multiple eventid x year observations 
. bysort eventid year: gen count = _N

. drop if count > 1 & postchange == 0
(2,537 observations deleted)

. drop count

. bysort eventid year: gen count = _N

. drop if count > 1
(129 observations deleted)

. drop count

. bysort eventid: egen min = min(postchange)

. bysort eventid: egen max = max(postchange)

. drop if min == max
(683 observations deleted)

. drop min max

. *** set globals
. global firm = "firmage firmage2"

. global ceo = "gender age age2 yrinco i.education"

. global eventsample = "moretrust == 1 | lesstrust == 1"

. global cluster = "eventid"

. keep if ${eventsample}
(1,018 observations deleted)

. 
. *** DE CHAISEMARTIN & D'HAULTFOEUILLE
. cap drop educ_*

. forval i = 1/4 {
  2.         gen educ_`i' = education == `i'
  3.         replace educ_`i' = . if education == .
  4. }
(1,959 real changes made, 1,959 to missing)
(1,959 real changes made, 1,959 to missing)
(1,959 real changes made, 1,959 to missing)
(1,959 real changes made, 1,959 to missing)

. * Col (1):
. did_multiplegt_stat ash_f1allpat eventid year trust_sd, ///
>         controls(firmage* gender age* yrinco educ_*) cluster(${cluster})
                                  ----------------------------------------------
                                   Number of observations     =            15448
                                   WAS Estimation method      =    doubly-robust
                                   Polynomial order           =                1
                                  ----------------------------------------------
                               (Std. err. adjusted for 2273 clusters in eventid)
--------------------------------------------------------------------------------
                               Average Slope (AS)
--------------------------------------------------------------------------------

             |  Estimate         SE      LB CI      UB CI  Switchers    Stayers 
-------------+-----------------------------------------------------------------
          AS | -2.748409   1.989783  -6.648383   1.151565       2223      13225 
--------------------------------------------------------------------------------
                           Weighted Average Slope (WAS)
--------------------------------------------------------------------------------

             |  Estimate         SE      LB CI      UB CI  Switchers    Stayers 
-------------+-----------------------------------------------------------------
         WAS |  .0513902   .0240398   .0042723   .0985082       2223      13225 
 
--------------------------------------------------------------------------------
--------------------------------------------------------------------------------

. 
. *** TWO-WAY FIXED EFFECTS
. * Col (2):
. reghdfe ash_f1allpat c.moretrust##c.postchange ${firm} ${ceo}, ///
>         a(eventid year) cluster(${cluster})
(dropped 45 singleton observations)
(MWFE estimator converged in 7 iterations)
note: moretrust is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)

HDFE Linear regression                            Number of obs   =     16,690
Absorbing 2 HDFE groups                           F(  11,   2168) =       2.47
Statistics robust to heteroskedasticity           Prob > F        =     0.0046
                                                  R-squared       =     0.9101
                                                  Adj R-squared   =     0.8965
                                                  Within R-sq.    =     0.0034
Number of clusters (eventid) =      2,169         Root MSE        =     0.5574

                                        (Std. err. adjusted for 2,169 clusters in eventid)
------------------------------------------------------------------------------------------
                         |               Robust
            ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------------------+----------------------------------------------------------------
               moretrust |          0  (omitted)
              postchange |  -.0797935   .0255645    -3.12   0.002     -.129927   -.0296601
                         |
c.moretrust#c.postchange |   .1030135   .0304885     3.38   0.001     .0432237    .1628033
                         |
                 firmage |  -.0142902   .0077361    -1.85   0.065    -.0294612    .0008808
                firmage2 |   .0000203   .0000862     0.24   0.814    -.0001487    .0001893
                  gender |  -.0278415   .0584232    -0.48   0.634    -.1424128    .0867297
                     age |   .0080572     .01141     0.71   0.480    -.0143184    .0304329
                    age2 |  -.0001037   .0001023    -1.01   0.311    -.0003044    .0000969
                  yrinco |   .0023375   .0014917     1.57   0.117    -.0005877    .0052627
                         |
               education |
                      2  |   .0789389    .051057     1.55   0.122    -.0211868    .1790646
                      3  |   .0838793   .0520759     1.61   0.107    -.0182445    .1860031
                      4  |   .1222121   .0567014     2.16   0.031     .0110173     .233407
                         |
                   _cons |   1.255072   .3553828     3.53   0.000     .5581454    1.951999
------------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     eventid |      2169        2169           0    *|
        year |        13           1          12     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. 
. *** SUN & ABRAHAM
. cap drop year_*

. forval i = 0/14 {
  2.         gen year_`i' = year == `i'      
  3. }

. drop year_6

. cap drop not*

. gen notmoretrust = !moretrust   

. gen notlesstrust = !lesstrust   

. * Col (3):
. eventstudyinteract ash_f1allpat postchange, cohort(cohort) control_cohort(notmoretrust) ///
>         covariates(${firm} ${ceo} postchange) absorb(eventid year) vce(cluster ${cluster})
(obs=8,308)

IW estimates for dynamic effects                        Number of obs = 16,690
Absorbing 2 HDFE groups                                 F(21, 2168)   =   1.76
                                                        Prob > F      = 0.0175
                                                        R-squared     = 0.9102
                                                        Adj R-squared = 0.8966
                                                        Root MSE      = 0.5572
                            (Std. err. adjusted for 2,169 clusters in eventid)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
  postchange |   .0981359    .031299     3.14   0.002     .0367568    .1595151
------------------------------------------------------------------------------

. * Col (4):      
. eventstudyinteract ash_f1allpat postchange, cohort(cohort) control_cohort(notlesstrust) ///
>         covariates(${firm} ${ceo} postchange) absorb(eventid year) vce(cluster ${cluster})
(obs=8,427)

IW estimates for dynamic effects                        Number of obs = 16,690
Absorbing 2 HDFE groups                                 F(21, 2168)   =   1.68
                                                        Prob > F      = 0.0270
                                                        R-squared     = 0.9101
                                                        Adj R-squared = 0.8965
                                                        Root MSE      = 0.5574
                            (Std. err. adjusted for 2,169 clusters in eventid)
------------------------------------------------------------------------------
             |               Robust
ash_f1allpat | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
  postchange |  -.1257225   .0321349    -3.91   0.000    -.1887409   -.0627041
------------------------------------------------------------------------------

. 
. *** CALLAWAY & SANT'ANNA
. cap drop cohort_*

. gen cohort_more = cohort

. replace cohort_more = 0 if !moretrust
(9,417 real changes made)

. gen cohort_less = cohort

. replace cohort_less = 0 if !lesstrust
(9,305 real changes made)

. * Col (5):
. csdid ash_f1allpat ${firm} ${ceo} postchange, ivar(eventid) time(year) gvar(cohort_more) agg(simple)
Panel is not balanced
Will use observations with Pair balanced (observed at t0 and t1)
..................................................
..................................................
................................
Difference-in-difference with Multiple Time Periods

                                                        Number of obs = 16,046
Outcome model  : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         ATT |   .1493422   .0414434     3.60   0.000     .0681146    .2305698
------------------------------------------------------------------------------
Control: Never Treated

See Callaway and Sant'Anna (2021) for details

. * Col (6):
. csdid ash_f1allpat ${firm} ${ceo} postchange, ivar(eventid) time(year) gvar(cohort_less) agg(simple)
Panel is not balanced
Will use observations with Pair balanced (observed at t0 and t1)
..................................................
..................................................
................................
Difference-in-difference with Multiple Time Periods

                                                        Number of obs = 16,098
Outcome model  : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         ATT |  -.1258161   .0411431    -3.06   0.002    -.2064551   -.0451771
------------------------------------------------------------------------------
Control: Never Treated

See Callaway and Sant'Anna (2021) for details

. 
. *** BORUSYAK ET AL.
. cap drop cohort_*

. gen cohort_more = cohort

. replace cohort_more = . if !moretrust
(9,417 real changes made, 9,417 to missing)

. gen cohort_less = cohort

. replace cohort_less = . if !lesstrust
(9,305 real changes made, 9,305 to missing)

. * Col (7): 
. did_imputation ash_f1allpat eventid year cohort_more, ///
>         controls(postchange firmage* gender age* yrinco educ_*) cluster(${cluster}) autosample tol(0.001)
Warning: part of the sample was dropped for the following coefficients because FE could not be imputed: tau.

                                                        Number of obs = 16,184
------------------------------------------------------------------------------
ash_f1allpat | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         tau |    .119013   .0311907     3.82   0.000     .0578804    .1801456
  postchange |  -.0673649   .0283557    -2.38   0.018    -.1229411   -.0117888
     firmage |  -.0084616   .0045017    -1.88   0.060    -.0172849    .0003616
    firmage2 |  -.0000392   .0001064    -0.37   0.713    -.0002477    .0001693
      gender |  -.0436222   .0713367    -0.61   0.541    -.1834395    .0961951
         age |   .0034439   .0162633     0.21   0.832    -.0284316    .0353194
        age2 |   -.000074   .0001458    -0.51   0.612    -.0003597    .0002118
      yrinco |   .0040602   .0020013     2.03   0.042     .0001377    .0079828
      educ_1 |          0  (omitted)
      educ_2 |   .0269221   .0625691     0.43   0.667     -.095711    .1495553
      educ_3 |   .0345523   .0650258     0.53   0.595    -.0928959    .1620004
      educ_4 |   .0996825   .0679868     1.47   0.143    -.0335691    .2329341
------------------------------------------------------------------------------

. * Col (8):
. did_imputation ash_f1allpat eventid year cohort_less, ///
>         controls(postchange firmage* gender age* yrinco educ_*) cluster(${cluster}) autosample tol(0.001)
Warning: part of the sample was dropped for the following coefficients because FE could not be imputed: tau.

                                                        Number of obs = 16,227
------------------------------------------------------------------------------
ash_f1allpat | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         tau |  -.1118135   .0337845    -3.31   0.001      -.17803    -.045597
  postchange |  -.0024307   .0306944    -0.08   0.937    -.0625907    .0577293
     firmage |  -.0813213   .0362921    -2.24   0.025    -.1524526   -.0101901
    firmage2 |   .0000599   .0001134     0.53   0.597    -.0001624    .0002822
      gender |  -.0082959   .0961469    -0.09   0.931    -.1967402    .1801485
         age |   .0087213   .0142748     0.61   0.541    -.0192568    .0366994
        age2 |  -.0000956   .0001278    -0.75   0.454    -.0003462    .0001549
      yrinco |   .0003335   .0022073     0.15   0.880    -.0039928    .0046598
      educ_1 |          0  (omitted)
      educ_2 |   .1284053   .0800379     1.60   0.109     -.028466    .2852766
      educ_3 |   .1304877   .0810189     1.61   0.107    -.0283065    .2892819
      educ_4 |   .1430154   .0901731     1.59   0.113    -.0337206    .3197514
------------------------------------------------------------------------------

. 
. cap log close
